A Petri net Semantic for BPEL4WS - Validation and Application

合集下载

业务流程管理BPM的认识(精)3完整篇.doc

业务流程管理BPM的认识(精)1第3页language such as provided by the Workﬂow Management Coalition are not precise enough.As always,it is preferable to identify any problems in software before it is actually deployed.In the case of Business Process Models this is especially im-portant as they may involve core business and/or complex business transactions.To reduce the risk of costly corrections,a thorough analysis of a Business Pro-cess Model can be beneﬁcial.Analysis of Business Process Models can also be used to investigate ways of improving processes(e.g.reducing their cost.Formal languages may have associated analysis techniques which can be used for inves-tigating properties of speciﬁcations.These techniques can then be relied upon to provide i nsight into the behavior and characteristics of a Business Process Model speciﬁed in such a language.In[1]three reasons are stated arguing the beneﬁts of the use of Petri nets for the speciﬁcation of workﬂows.The reasons brought forward are the fact that Petri nets are formal,have associated analysis techniques,and are state-based rather than event-based.The development of Woﬂan(see e.g.[25]demonstrates that workﬂows speciﬁed as workﬂow nets[2],a subclass of Petri nets,can be analyzed in order to determine whether they are e.g.deadlock free.In the context of UML activity diagrams,tool support forveriﬁcation is discussed in[15].Through the notion of place,Petri nets provide natural support for modeling the stages in between processing.State-based patterns such as deferred choice, interleaved paralled routing,and milestone can therefore be speciﬁed straightfor-wardly.A description of these patterns can be found in[9].Petri nets though also have some deﬁciencies when it comes to the speciﬁcation of certain controlﬂow dependencies(see[7].This observation has led to the development of YAWL[8] (Yet Another Workﬂow Languageof which the formal semantics is speciﬁed as a transition system.It is interesting to observe that a concept such as the deferred choice,while easily captured in terms of Petri nets,is not often supported in languages of “classical”workﬂow management systems(see[9].Two recently proposed stan-dards for web service composition,BPEL4WS and BPML,however,provide direct support for thisconstruct(see[27]and[5]resp..In web services compo-sition it is important to capture the interactions between the various services and a formalism such as theπ-calculus seems to be a natural candidate to provide a formal foundation for such interactions.While it is some times claimed that BPML is based on theπ-calculus,there does not seem to be a precise deﬁnition of this relation(note that in[12],it is stated that“there is currently no evidence that BPEL4WS is based on a formal semantics”.We believe that it is important that such relations are fully formalized.Formally deﬁned Business Process Modeling Languages can be compared in terms of their expressive power.For some classes of workﬂow modeling languages, abstractions of some existing approaches,comparative expressiveness has been studiedin[21,20].These results are in the context of a speciﬁc notion of equiv-alence,addressing the issue of when two workﬂow models can be considered expressing the sameworkﬂow.Expressiveness results give insight into what can and cannot be expressed in some approaches and more research is needed in this area as it could provide more guidance for language development.4Available Technology and Emerging StandardsBased on the deﬁnition of Business Process Management proposed in Section2, a characterization of its main concepts is provided,and the technology currently available or on the horizon is discussed.Some of the key aspects of business process management already mentioned in Sections1and2are re-visited,and the current state of available technology and emerging standards are discussed.One of the main aspects and certainly an activity typically carried out in early phases of business process management projects is the design of business pro-cesses.There is a close relationship between business process design and business process modeling,where the former refers to the overall design process involving multiple steps and the latter refers to the actualrepresentation of the business process in terms of a business process model using a process language.To this end,the term business process modeling is used to characterize the identiﬁca-tion and(typically rather informalspeciﬁcation of the business processes at hand.This phase includes modeling of activities and their causal and temporal rela tionships as well as speciﬁc business rules that process executions have to comply with.Business process modeling has a decade long tradition,and a variety of prod-ucts are commercially available to support this phase,based on diﬀerent process languages.Given this situation,it is not surprising that the selection of a par-ticular product is an important step in many BPM projects,and,consequently, appropriate selection criteria have been studied extensively.Besides organiza-tional,economical,and aspects related to the overall IT infrastructure of the enterprise at hand,the expressive power of the process language as well as inter-faces to related software systems are important criteria,most prominently inter-faces to process enactment systems(such as workﬂow management systemsand to software responsible for modeling personnel and organizational structures of the enterprise.Not only the expressive power but also a well-deﬁned semantics of the process language deserves a central role during product selection.How-ever,this aspect is considered only in a small number of recent business process management projects.Business process analysis aims at investigating properties ofbusiness pro-cesses that are neither obvious nor trivial.To this end,the term analysis is used with a rather broad meaning,including for example simulation and diagnosis, veriﬁcation and performance analysis.Process simulation facilitates process di-agnosis in the sense that by simulating real-world cases,domain experts can acknowledge correct modeling or proposemodiﬁcations of the original process model.If business process models are expressed in process languages with a clear semantics,their structural properties can be analyzed.If,for example,certain parts of processes can never be reached,an obvious modeling mistake occurred that should beﬁxed.While basic structural properties of process models have been studied for some time,it is remarkable that few software products actually support them.However structural analysis of process models requires a clear for-mal semantics of the underlying process language,which might not be present. In some products,a pragmatic approach to process modeling is preferred to aformal one;especially if the main goal of process modeling is discussion with domain experts rather than process analysis or process enactment.However,we mention that formal semantics of process languages and intuitiveness and ease of use are no contradicting goals,and recent approaches seem to support this observation.The next aspect of BPM and traditionally a very strong one is process en-actment.However,before process enactment isdiscussed,we provide a coarseclassiﬁcation of business processes that paves the way for a discussion of dif-ferent types of process enactment systems.In the early days of BPM when in the application side business process modeling and in the IT enactment side workﬂow management were the only options,processes with a static structure were focused.The main reason behind this obvious limitation was as follows: Modeling a process and providing infrastructure for its enactment incurs con-siderable eﬀort.To provide satisfactory return on investment,a large number of individual cases have to beneﬁt from this new technology.This type of straight-through-process is also ca lled production workﬂow[23].While there are success-ful workﬂow projects on this type of straight-through processes,this restriction ofworkﬂow technology proved fatal for applications in more dynamic environ-ments.In some cases where traditional workﬂow t echnology was used in these advanced settings,new workﬂow solutions were partly circumvented or even ne-glected.As a response to this situation,considerable work in ad-hoc,ﬂexible and case-based workﬂowwas(and is beingconducted,both in academia and in industry.Recently,case handling is studied in depth as a new paradigm for supporting knowledge-intensive business processes with loose structuring.Based on the brief characterization of case handling provided above,we mention that in the case handling paradigm knowledge workers enjoy a great degree of freedom inorganizing and performing their work which they are knowledgeable about. Some of the concepts of case handling are already present in commercial case handling systems.Standardization has a long histor y in workﬂow management.Fueled by infor-mation system heterogeneity that also includes workﬂow management systems, organizations started to form interest groups aiming at standardizing interfaces between workﬂow management systems and components,with the goal of en-hancing interoperability and fostering the workﬂow market.The most prominent organization in this context is the Workﬂow Management Coalition(WfMCthat was formed in1993and today hasover300member organizations,including all major workﬂow vendors as well as workﬂow users and interested academia[22]. The basis of WfMC activities is the so called WfMC Reference Architecture that deﬁnes standard workﬂow system components interfaces.Despite the fact that all major vendors are organized in WfMC and a number of important contri-butions on practical workﬂow aspects have been made,many people feel that WfMC’s ambitious goals have yet to be reached.A more recent standardization eﬀort in the BPM context is related to the cur-rent momentum of XML and Web services technology.Web services is a promis-ing technology to foster interoperability between information system based–conceptually–on the service oriented architectureparadigm[11]and–tech-nologically–on open standards and light-weight protocols and systems.While Web services technology has not yet reached maturity level,there is considerable eﬀort under way by literally all major software vendors.The need for standard-ization is clearly acknowledged in this context,and important contributions have been made.However,as。

共享合成Petri网系统的语言递归性与系统活性

1
基本概念
关于 Petri 网系统的基本概念及术语可参见文献[10~12],这里只给出本文需要的一些概念. 定义 1.1[10~12]. Σ=(P,T;F,M0)是一个 Petri 网系统,称 Ls(Σ)={α|α∈T*且 M0[α>}为网系统 Σ的顺序语言或网系
统Σ的顺序行为,其中 T*是 T 的闭包.
蒲飞等:共享合成 Petri 网系统的语言递归性与系统活性
319
定义 1.2[10~12]. Σ=(P,T;F,M0)是一个 Petri 网系统,称 Lc(Σ)={α|α∈(P(T*))*且 M0[α>],为网系统 Σ的并发语言或
Σ的并发行为,其中 P(T*)表示 T 的幂集.
通常,在 Rt∈P(T*)时,|Rt|=0 称为空步,记为 Rt=ε,|Rt|=1 称为单步,|Rt|≥2 称为并发步,其中 |Rt|表示 Rt 含有的字符串个数. 定义 1.3. 设 L 是Σ上的一个并发语言,语句α,β∈L,则称 “◦”为 L 的乘法运算,意为α◦β≡αβ,是一种并发串的连接.称“+”为 L 的加法运算,意为α+β≡{α}+{β}={α,β}. 我们很关心系统的活性和无死锁性 , 因为系统具有活性意味着在任何状态下系统的任意一个功能部件均有机会产生作用 ; 而系统具有无死锁性意味着系统在任何状态下还会往前运行 . 这些性质都是十分重要的 . 定义 1.4[10~12]. 设 Petri 网系统 Σ=(P,T;F,M0),如果∀M∈[M0>,∀t∈T,∃M′∈[M>,使 M′[t>,则称 Petri 网系统 Σ是活的.如果对∀M∈[M0>,∃t∈T,使得 M[t>,则称 Petri 网系统Σ无死锁. 定义 1.5[10~12]. 设 Petri 网系统 Σ i = ( Pi , Ti ; Fi , M 0i ) (i=1,2).令Σ=(P,T;F,M0),使得 1) P= P 1∪P 2 , P 1∩P 2 ≠∅ ; 2) T= T1 ∪ T2 , T1 ∩ T2 = ∅ ; 3) F= F1 ∪ F2 ; max{M 01 ( p), M 02 ( P)}, 若p ∈ P1 ∩ P2 , 4) M 0 ( p ) = 若p ∈ Pi − ( P1 ∩ P2 ); (i = 1,2) M 0i ( p ),

人工智能控制技术课件：神经网络控制

进行的，这种排列往往反映所感受的外部刺激的某些物理特征。
例如，在听觉系统中，神经细胞和纤维是按照其最敏感的频率分
布而排列的。为此，柯赫仑（Kohonen）认为，神经网络在接受外
界输入时，将会分成不同的区域，不同的区域对不同的模式具有
不同的响应特征，即不同的神经元以最佳方式响应不同性质的信
号激励，从而形成一种拓扑意义上的有序图。这种有序图也称之

,

,
⋯
,

)
若输入向量 X= ( 1
，权值向量
2

W=(1 , 2 , ⋯ , ) ，定义网络神经元期望输出与
实际输出的偏差E为：
E= −
PERCEPTRON学习规则
感知器采用符号函数作为转移函数，当实际输出符合期
望时，不对权值进行调整，否则按照下式对其权值进行
单神经元网络
对生物神经元的结构和功能进行抽象和
模拟，从数学角度抽象模拟得到单神经
元模型，其中是神经元的输入信号，
表示一个神经元同时接收多个外部刺激；
是每个输入所对应的权重，它对应
于每个输入特征，表示其重要程度；
是神经元的内部状态；是外部输入信
号；是一个阈值（Threshold）或称为
第三代神经网络：
2006年，辛顿（Geofrey Hinton）提出了一种深层网络模型——深度
置信网络（Deep Belief Networks，DBN），令神经网络进入了深度
学习大发展的时期。深度学习是机器学习研究中的新领域，采用无
监督训练方法达到模仿人脑的机制来处理文本、图像等数据的目的。
控制方式，通过神经元及其相互连接的权值，逼近系统

Knockout of UBP34 in Physcomitrella patens reveals the photoaffinity labeling of another

Plant Science 167(2004)471–479Knockout of UBP34in Physcomitrella patens reveals the photoafﬁnitylabeling of another closely related IPR proteinFlorent Brun a ,Didier G.Schaefer b ,Michel Laloue a ,Martine Gonneau a ,∗aLaboratoire de Biologie Cellulaire,INRA-Versailles,route de Saint-Cyr,78026Versailles-Cedex,Franceb Laboratoire de Phytogénétique Cellulaire-Institut d’Écologie,Universitéde Lausanne,Bˆa timent de Biologie,CH-1015Lausanne-Dorigny,Switzerland Received 15September 2003;received in revised form 15April 2004;accepted 20April 2004Available online 18May 2004AbstractUBP34,a soluble 34kDa protein identiﬁed in the moss Physcomitrella patens by photoafﬁnity labeling with an urea-type cytokinin agonist,belonging to the plant intracellular pathogenesis related (IPR)protein family.This class of proteins is ubiquitous in the plant kingdom but their functions are still ing a reverse genetic approach we generated knockout mutants at the UBP34locus by homologous recombination.Knockout plants grown in various conditions do not present any visible phenotypic alteration.However,biochemical and molecular analysis of the knockout transformants reveal the speciﬁc photoafﬁnity labeling of another similar protein.©2004Elsevier Ireland Ltd.All rights reserved.Keywords:Allergen;Gene disruption;Intracellular PR proteins;Physcomitrella patens1.IntroductionUsing the cytokinin agonist azido-CPPU (azido-1-(2-chloropyrid-4-yl)-3-phenylurea)we speciﬁcally pho-toafﬁnity labeled a 34kDa soluble protein in the moss Physcomitrella patens [1].The protein was puriﬁed by afﬁnity chromatography,micro-sequenced,and the corre-sponding cDNA was cloned by screening of a P .patens cDNA library.This protein,named UBP34for urea-type CK-binding protein,is homologous to plant intracellular pathogenesis related (IPR)proteins,known as PR10pro-teins [2].UBP34is twice the usual size of IPR proteins and consists of two tandemly arranged PR10units,the highest identity being located in the N-terminal half part of the pro-tein [1].PR10proteins were originally characterized at the transcription level,demonstrating increased gene expression in stress situations,such as during microorganism infection.Several PR10proteins have been described as plant aller-Abbreviations:IPR,intracellular pathogenesis related;IP,isopenteny-ladenine;CPPU,1-(2-chloropyrid-4-yl)-3-phenylurea;BA,benzyladenine ∗Corresponding author.Tel.:+33-1-30833046;fax:+33-1-30833099.E-mail address:gonneau@versailles.inra.fr (M.Gonneau).gens.Among them the Bet V 1protein in Betula verrucosa was recently shown to bind a broad spectrum of physiologi-cal ligands including fatty acids,ﬂavonoids,and cytokinins (isopentenyladenine and kinetin)[3].Other proteins of this family have also been described as cytokinin-binding proteins in Vigna radiata [4]or as protein,of which the cor-responding gene is transcriptionally activated by cytokinin treatments in Catharanthus roseus [5].However,a role of UBP34strictly in term of cytokinin perception can be ruled out with regard to previous pharmacological data indicat-ing that N 6-substituted adenine derivatives are poor ligands for this protein [1].The functional role of the IPR protein family remains to be established as well as the physiolog-ically signiﬁcant in their cytokinin regulation and binding properties.In the moss P .patens ,gene knockout,and allele re-placement approaches are feasible as was illustrated by the disruption of various genes involved in different aspects of plant physiology (for a review,see [6]).We aimed to take advantage of this unique property in the plant kingdom to specify the physiological function of the UBP34protein in moss.Therefore,we cloned the UBP34gene,obtained the corresponding knockout mutants0168-9452/$–see front matter ©2004Elsevier Ireland Ltd.All rights reserved.doi:10.1016/j.plantsci.2004.04.013472F.Brun et al./Plant Science167(2004)471–479and analyzed their phenotypes.Under the different growth conditions that we have tested,knockout plants do not show developmental or morphological alteration.However,bio-chemical and molecular analysis of the knockout mosses reveals the labelling with the azido-CPPU photoafﬁnity probe of another closely related IPR protein.This suggests a possible phenomenon of functional compensation due to redundancy at the genomic level.In the UBP34-knockout plants,a regulatory mechanism of both proteins could be modiﬁed and affect the protein UBP34-like stability and/or activity.2.Materials and methods2.1.Plant material and culture conditionsThe Gransden Wild-type strain of P.patens and growth conditions have been described previously[1].2.2.Analysis of nucleic acids and genomic library screeningDNA manipulations were performed according to stan-dard procedures[7].The genomic library of P.patens in the bacteriophage lambdaGEM12was kindly provided by M. Leach(John Innes Institute,UK).The library was built by ligation of Sau IIIA partially digested DNA fragments at the Xho I site of the lambda phage.The library was infected into Escherichia coli host strain LE392.Immobilization of DNA on Hybond–N membrane(Amersham),probe hybridization with UBP34cDNA,and isolation of phages of interest were performed according to[7].2.3.Targeting vector constructionA2,6kb Eco RI fragment was isolated from the lambda phage containing a5.5kb fragment of the UBP34locus genomic sequence.This fragment was subcloned into the Eco RI site of a modiﬁed pBSII–KS−plasmid(Stratagene)in which the Xho I site of the multiple cloning site has been re-moved.For disruption experiments of the UBP34locus,the replacement vector pUBP34– XhoI–NptII was constructed by excision of the genomic internal Xho I fragment and re-placement by the Npt II selectable marker gene under control of the35S promoter and terminator of the cauliﬂower mo-saic virus[8].Constructions containing the marker gene in the same orientation as the UBP34gene were selected by PCR.2.4.Protoplast isolation and transformationFor protoplast isolation,protonema tissues were propa-gated on the minimal medium described by[9]and supple-mented with2.7mM NH4tartrate(Pp-NH4medium)and 25mM glucose.Cultures were grown on cellophane disks in9cm Petri dishes on medium solidiﬁed with0.7%Agar (Biomar)at25◦C,with a light regime of16h light/8h dark-ness and a quantum irradiance of80␮E m−2s−1.Protoplasts were isolated from6-day-old protonemal cultures as de-scribed by[8].Transformation experiments were performed with300␮l(4×105protoplasts)of a protoplast suspension added to30␮l of Eco RI digested pUBP34– XhoI–NptII plasmid(15␮g)according to the classical protocol[8]. The regenerating colonies were submitted to the paro-momycine(DUCHEFA)selection pressure(50␮g ml−1) for7days.Stable resistant clones were deﬁnitely selected after a second round of growth on non-selective Pp-NH4 medium followed by transfer on Pp-NH4medium containing paromomycine.2.5.DNA and RNA extractionsP.patens DNA was extracted according to a method de-scribed by[10].Brieﬂy,200mg of fresh tissue were ground in liquid nitrogen,mixed with600␮l extraction buffer (100mM,Tris–HCl pH8;50mM,EDTA;500mM,NaCl; 10mM,␤-mercaptoethanol),and50␮l20%SDS and in-cubated at65◦C for10min.230␮l5M potassium acetate were then added.The mixture was kept on ice for20min and centrifuged at16,000×g for20min.DNA was then extracted twice with phenol–chloroform–isoamylalcohol and precipitated with1.5volume of isopropanol.The sam-ple was then treated with RNAse-A,extracted again with phenol–chloroform–isoamylalcohol,precipitated in abso-lute ethanol with3M sodium acetate and dissolved in25␮l Tris–EDTA,pH8.Total RNA from P.patens protonema tissues were extracted using the Qiagen RNeasy Plant extraction kit.2.6.Southern and Northern analysisFor Southern blot analysis,DNA(10␮g)was digested with the appropriate restriction enzymes.Fragments were separated by electrophoresis in0.8%(w/v)agarose–Tris Ac-etate EDTA(TAE)-gels and transferred to nylon membranes (GeneScreen Plus,NEN)in20×SSC(3M NaCl,0.3M sodium citrate,pH7).For Northern blot analysis,7␮g of total RNA were sepa-rated on1.0%(w/v)agarose gels containing formaldehyde–MOPS and transferred to nylon membranes(GeneScreen Plus,NEN)in20×SSC(3M NaCl,0.3M sodium citrate, pH7).Probes were radiolabelled with30␮Ci[␣32P]dCTP by random priming(Prime-a-Gene labeling system,Promega). Hybridisation was conducted in modiﬁed Church buffer (7%SDS;0.25-Na2HPO4,pH7.4;0.2mg/ml heparin; 2mM EDTA;0.1mg ml−1denatured DNA of salmon sperm)for16h at65◦C(high stringency)or55◦C(low stringency).Theﬁnal washes were in2×SSC and then0.2×SSC supplemented with0.1%SDS,at65◦C or55◦C, respectively.F.Brun et al./Plant Science167(2004)471–4794732.7.RT-PCR and PCR analysisThe PCR reaction mixture(20␮l)contained200␮M of each dNTP,500nM of each primer,1.5mM MgCl2,and one unit of Taq polymerase in PCR buffer(Gibco BRL).The DNA was denatured at94◦C for3min.Ampliﬁcation was performed for35cycles of30s denaturation at94◦C,30s annealing at55◦C,and1min elongation at72◦C,followed by a10minﬁnal elongation step at72◦C.10␮l of the re-action mixture were analyzed on a1%TAE–agarose gel. Speciﬁc primers used for the characterization of the UBP34genomic locus of the transformant lines were:U5 A(5 -CGTCTAGGGATACAGGAAGG-3 ),L3 A(5 -GACTACATACTGCACCATTG-3 ),Uc(5 -TATTTTTGG-AGTAGACAAGCGTGTCGT-3 ),UC2(5 -TAATGTGTG-AGTAGTTCCCAGATAAGG-3 ).For RT–PCR,double-stranded cDNAs were prepared from wild type total RNA by using the SMART–PCR cDNA Synthesis Kit(Clontech).Speciﬁc primers used for the ampliﬁcation of the UBP34-like partial cDNA frag-ment were:U171(5 -CATTGATCCCCTCCCTCCC-3 )and L173(5 -TCCCCCGAACACAAAACCC-3 ).2.8.Sequence analysisSequence data analysis were performed by Blast search at NCBI against the P.patens EST database(http://www. /blast/;[11])and at the NIBB PHYSCObase (http://www.moss.nibb.ac.jp/).2.9.Protein extraction,photoafﬁnity labeling,and afﬁnity chromatographyProteins from P.patens wild type and UBP34-knockout transformants were extracted and photoafﬁnity labeled as previously described[1].Puriﬁcation of the UBP34-like protein on NH2CPPU–Sepharose6B afﬁnity column was achieved as previously described[1].2.10.Mass spectrometry analysisAfﬁnity puriﬁed proteins were analyzed by SDS-PAGE. After staining with Coomassie blue(Bio-Safe Coomassie stain,Bio-Rad),bands were excised and washed twice in25mM ammonium carbonate(pH8),for30min,then twice in50%acetonitrile,25mM ammonium carbonate for 30min andﬁnally with100%acetonitrile and dehydrated in a speed-vacuum dryer.The cysteine residues were reduced by45min treatment at56◦C,with10mM DTT in100mM ammonium car-bonate(pH8).The reduced cysteine residues were then alkylated by55mM iodoacetamide in100mM ammonium carbonate(pH8),30min in the dark,under vigourous agi-tation.The gel pieces were then washed twice with100mM ammonium carbonate(pH8),acetonitrile(50/50;v/v)and speed-vacuum dried.The gel pieces were re-swollen with 0.25␮g of sequence grade modiﬁed bovine trypsin(Roche)in25mM ammonium carbonate,pH8until the gel wascompletely re-hydrated and incubated at37◦C for18h.The tryptic supernatant was then transferred to a cleanEppendorf tube and the gel pieces were covered with0.1%TFA in60%acetonitrile and sonicated for15min.This op-eration was repeated twice and the two supernatants werepooled with theﬁrst tryptic supernatant.The supernatantswere speed-vacuum concentrated.MALDI-TOF mass spectra were recorded under the re-ﬂectron mode with a Reﬂex-III(Bruker SA,Karlsruhe,Ger-many)equipped with a nitrogen laser operating at337nm(3ns pulse duration).The accelerating voltage used was16kV and the delay time setting was200ns.Each spectrumwas produced by accumulating data from100–250lasershots.Mass spectra were calibrated with an external stan-dard containingﬁve peptides in the1000–2500m/z range.The matrix solution was prepared by dissolving10mg of ␣-cyano-4-hydroxycinnamic acid in500␮L distilled water, 500␮L acetonitrile,and1␮L triﬂuoroacetic acid.One mi-croliter of this solution was placed on the sample plate to-gether with1␮L of the sample solution and allowed to dryat room temperature.The sample plate was then placed inthe instrument.3.Results3.1.Characterization of the UBP34genomic locus andSouthern blot analysisA P.patens genomic library was screened by hybridiza-tion using the full length UBP34cDNA as a probe.A pos-itive clone was puriﬁed and a5.5kb Eco RV fragment wassubcloned in pBS–KS−and sequenced.This clone con-tains the entire cDNA sequence plus1424bp in5 and1050bp in3 (Fig.1A).The genomic DNA includes twointrons of295and132bp.A potential site for initiationof the transcription is located128bp upstream of the ATGcodon(AATATATAGA).The TAA stop codon is located1297bp downstream of the ATG codon on the genomic DNA(Fig.1A).P.patens genomic DNA was analyzed by Southern blothybridisation with the internal cDNA Xho I fragment as aprobe(Fig.1A).Under high stringency conditions(65◦C),the hybridization pattern indicated a unique UBP34copyin the moss genome,with two locus-speciﬁc fragments of2.6and3.3kb,obtained with Eco RI and with Hin dIII,re-spectively(Fig.2A).The double digestion with Hin dIIIand Xho I produced the expected1.2kb fragment(Fig.2A).When hybridization was performed at55◦C,other weakhybridization signals appear(Fig.2B)indicating that thegenome includes sequences homologous to UBP34.Indeed,several overlapping P.patens EST sequences which are ho-mologous to UBP34can be detected in the Genbank ESTdatabase.The resulting EST contig encodes a290amino474F .Brun et al./Plant Science 167(2004)471–479Fig.1.Structure of the UBP34locus (A)and of the replacement vector pUBP34– XhoI–NptII and (B)used for gene disruption.The two introns of the UBP34gene are represented by light gray boxes.Location of primers used to screen for targeted transformants is shown by arrows.Hatched box represents the probe used for Southern analysis.(Eco RV site on the left belongs to the phage arm).acids protein which will be referred as UBP34-like (Fig.3).UBP34and UBP34-like share 76%identity at the amino acid level with each other.The alignment of a duplicated tan-dem arrangement of Pinus monticola PR10protein,one of the most closely related higher plant IPR,with UBP34and UBP34-like (Fig.3)emphasizes the imperfect duplication of the P .patens IPR proteins amino acid sequences [1].A cDNA fragment,covering all the open reading frame ofthe Fig.2.Southern analysis of wild type (A,B and D)and UBP34-knockout plants (C).Hybridization at high stringency (65◦C,A)or at low stringency (55◦C,B–D).Hybridization with the UBP34Xho I probe (A–C)and the UBP34-like cDNA probe (D).Restriction digest:H:Hin dIII;E:Eco RI,and X:Xho I.Arrows indicate signals detected at low stringency,with one probe or the other,which can be attributed neither to the UBP34locus nor to the UBP34-like locus.UBP34-like gene,was ampliﬁed by RT–PCR using speciﬁc primers and wild-type total RNA.This PCR fragment was used as a probe for low stringency Southern blot (Fig.2D ).It appears that two weak signals,of 2and 2.2kb,resulting from Hin dIII digestion,and detected at low stringency with the UBP34probe (Fig.2B ),correspond to the UBP34-like locus.However,other weak signals can not be attributed either to the UBP34or to the UBP34-like genes (Fig.2BF.Brun et al./Plant Science167(2004)471–479475Fig.3.Alignment of UBP34,UBP34-like peptide sequence(deduced from a contig of overlapping EST(BJ180871,BJ186774,BJ178893,BI487420, A W599666,BJ170059,BJ175657))and a duplicated tandem arrangement of Pinus monticola PR10protein,one of the most closely related higher plant IPR(AAL5006,from valine(V)11to glutamic acid(E)150).Alignment was achieved with the ClustalW program.and D),indicating that at least another homologous gene exists in the P.patens genome.A phylogenetic analysis was performed using14IPR related proteins:P43185the major birch pollen allergen (Betula pendula);P52779.1from yellow lupine which crystal structure has been established[12],the Arabidop-sis peptide At1g24020with the highest identity to IPR proteins;P80890from Panax ginseng which possess puta-tive RNase activity;AAB09084.1a17kDa isolated from Asparagus,a monocotyledonous species and induced by thiocarbamate[13];AAD26553from Malus domestica,this protein has been photofﬁnity labelled using the azido-CPPU probe(P.Simoneau,personal communication);P19418, PcPR1from Petroselium crispum theﬁrst isolated PR-10 protein;T10059a protein from Madagascar periwinckle induced by cytokinin treatment;BAA74451from Vigna ra-diata a cytokinin-speciﬁc binding protein;AAF12810and AAL50006,IPR-like from conifers(Picea glauca,Pinus monticola);BJ197020a single EST from P.patens,the ﬁrst150N-terminal aa of UBP34and UBP34-like proteins where the highest identity with IPR proteins is located (UBP34-like protein is also twice the size of higher plants IPR protein).The resulting tree indicates that BetV1and MalD1are very narrow and that UBP34and UBP34-like genes do not cluster away from IPR of higher plants species (Fig.4).No IPR homologous sequences for rice or maize are currently available.3.2.Molecular analysis of UBP34-knockout transformants The replacement vector consists of a3.2kb Eco RI frag-ment with two750bp and760bp genomic fragmentsﬂank-ing the Npt II cassette which replaces a1200bp genomic Xho I fragment(Fig.1B).Transformation experiments per-formed according to the classical method resulted in2×106regenerating protoplasts.Independent and stable trans-formed lines have been isolated with a relative transforma-tion frequency of4.6×10−3.First,75stable transgenic lines were screened by PCR for targeted integration of the replacement vector at the UBP34locus.PCR analysis with primers U5 A/UC and L3 A/UC2reveals30%of gene re-placement events among the75stable transformants.The sequences of the ampliﬁed fragments are in agreement with the corresponding sequences of the pUBP34– XhoI–NptII resulting from the replacement of the genomic Xho I frag-ment by the Npt II cassette.Nine knockout transformants were further characterized by Southern blot analysis.The restriction digests of genomic DNA using the Xho I fragment as a probe,show that the wild type major signals are lost in the knockout line(Fig.2C).However,the weak signals visi-ble only under low stringency conditions and corresponding to UBP34-like and other homolog(s),are identical in wild type and knockout lines.In Northern blot analysis with the Xho I probe,the UBP34 transcript of approximately1.2kb detected in wild type is absent in the knockout transformant,conﬁrming that a null mutant for the UBP34gene was obtained(Fig.5A-1).More-over the UBP34transcript does not appear up-regulated upon IP treatment(Fig.5A-1).3.3.Phenotypic analysis of the knockout transformants Knockout transformants do not exhibit visible morpho-logical alterations throughout the developmental cycle com-pared to wild type under standard culture conditions(liquid or solid medium).At that time,we could not visualize any overt difference in term of color or density of the tissues, growth rate,and number of gametophore on a protonema476F .Brun et al./Plant Science 167(2004)471–479Fig.4.Distance analysis of the IPR family using the full length amino-acid sequence of 14IPR proteins,the ﬁrst 150N-terminal aa of UBP34and UBP34-like proteins and the sequence of another related P patens protein corresponding to a single EST (Arabidopsis,At1g24020;Lupinus luteus ,P52779;Panax ginseng ,P80890;Betula pendula ,P43185;Asparagus ofﬁcinalis ,AAB09084.1;Malus domestica ,AAD26553;Petroselium crispum ,P19418;Catharanthus roseus ,T10059;Vigna radiata ,BAA74451;Picea glauca ,AAF12810;Pinus monticola ,AAL50006;Physcomitrella patens BJ197020(single EST)).The sequences were aligned using ClustalX under Phylip format and the phylogenetic tree was drawn using the Treeview software.base,etc.In order to reveal conditional phenotypes related to the UBP34loss-of-function,we tested the growth of the transformant under different conditions.UBP34was initially identiﬁed as a urea-type cytokinin binding protein.Therefore,the response to cytokininof Fig.5.Northern analysis of wild type (WT)and UBP34-knockout plants (KO).Total RNA from wild-type plants (treated or not with 0.2␮M IP for 5h)and from KO transformants were hybridized with the UBP34Xho I probe (A-1)or UBP34-like cDNA probe (A-2).As control,18S rRNA,with ethidium bromide (B).UBP34-knockout mosses was investigated.Moss transfor-mants resulting from germination of spores had similar phenotypes compared to wild type when grown on medium supplemented with 0.2␮M isopentenyladenine (IP).The response to urea type cytokinin was compared to theF .Brun et al./Plant Science 167(2004)471–479477response to adenine type cytokinin on 6-day-old vertically dark-grown protonema.Dark-grown mosses are negatively gravitropic [14],and such growth conditions offer the ad-vantage to obtain many unidirectional ﬁlaments.Tissues were treated with 0,30,and 100nM benzyladenine (BA)or CPPU for 72h and buds formed on caulonemata ﬁles were counted.The cytokinin-induced bud formation of knockout lines was similar to that of wild type.Considering that UBP34belongs to a family of proteins involved in stress responses in higher plants,the growth of wild type and knockout plants was compared under osmotic (50and 350mM mannitol)or saline (50and 350mM NaCl)stress situations.No visible difference was observed in these conditions between wild type and UBP34-knockout trans-formants.High salt concentrations were toxic and killed the colonies and 350mM mannitol resulted in reduced growth.3.4.Photoafﬁnity labeling and puriﬁcation of the UBP34-like protein in the knockout linesPhotoafﬁnity labeling of soluble proteins from wild type and knockout lines conﬁrmed the absence of the signal in the knockout lines.However,a new photoafﬁnity labeled protein with an apparent molecular weight of 36kDaap-Fig.7.MALDI-TOF peptide mass ﬁngerprint spectrum of a peptide mixture from in-gel tryptic digestion of the 36kDa photolabelled pro-tein.Eight peptide ions matched with predicted peptide ions masses expected from the tryptic digestion of UBP34-like protein (Peptide Mass;/cgi-bin/peptide-mass.pl )and covering 41%of the protein sequence;their theorical masses (m/z)are:1388.7824(VLPDLLPEF-FAK),1414.7172(TEILEGDGGPGTLR),1476.8595(VLHFGPAIPQAGAAK),2298.0805(LDTVDDATMTLSYTVVEGDPR),1657.8431(YVNVT-GVVSFASTGEK),1496.7379(YDVVGEAGPPEHVK),937.5175(NITALMFK),and 2114.0400(TATHTETLDASPDAIWSA VK).Peptide ions 1856.0827(HSDKVLPDLLPEFFAK),2583.2427(ERLDTVDDATMTLSYTVVEGDPR),and 2573.2758(MGPAIPDAGELVEQVDVFDDAEKK)result from one missed cleavage.Two peptides ions are identiﬁed as a result from trypsin (T)autolytic digestion (2163.1127and2273.1944).Fig.6.Photoafﬁnity labeling with [3H]azido-CPPU of soluble protein ex-tracts of wild type (WT)and UBP34-knockout (KO)P .patens protonema (25␮g of protein are loaded in each lane).Molecular weights (kDa)of the size markers are indicated on the left.pears in the knockout (Fig.6).This protein was puriﬁed on NH 2CPPU–Sepharose 6B afﬁnity column using soluble proteins from a knockout line,as previously described for purifying the UBP34protein [1].Thereafter MALDI-TOF peptide mass ﬁngerprint was realized after tryptic digestion of the 36kDa puriﬁed protein.Eight peptide ions matched masses of predicted peptides resulting from tryptic digestion478F.Brun et al./Plant Science167(2004)471–479of the UBP34-like protein,covering41%of this protein (Fig.7).The predicted molecular weights for the two proteins are very similar(UBP34:31103.23and UBP34-like:31115.31) and slightly distant from the apparent molecular mass.It may be related to different post-translational modiﬁcations and/or to peculiar three-dimensional folding.The UBP34-like cDNA fragment,previously used for Southern blot,was used as a probe for Northern blot anal-ysis of total RNA in both wild-type and knockout lines.A corresponding transcript of1.5kb,in good correlation with the predicted cDNA size,is present at the same level in wild type and in knockout lines(Fig.5A-2).The light signal ob-served on the Northern blot of KO transformants(Fig.5A-1) most probably corresponds to a cross hybridisation of the UBP34-probe with the UBP34-like transcript.4.DiscussionIn the moss P.patens,we identiﬁed a soluble34kDa protein that was speciﬁcally labeled by photoafﬁnity with the phenylurea-derived cytokinin agonist,azido-CPPU.The UBP34protein is homologous to plant IPR of unknown function.The UBP34gene belongs to a multigene family, among which the UBP34-like gene could be identiﬁed in the EST database.The open reading frames of UBP34and UBP34-like share76%identity at the amino acid level with each other.We took advantage of the powerful property of P. patens to knockout genes by homologous recombination,to analyze the phenotype of UBP34loss of function mutants. In P.patens,the requirement for strict sequence homology in gene targeting experiment was illustrated with the highly conserved chlorophyll a/b binding(cab)protein multigene family.Although sequence homology between11members of the cab multigene family is as high as86–94%at the nucleotide level,successful targeting of a speciﬁc member, the ZLAB1locus,has occurred in30%of the transgenic plants analyzed[15].The UBP34locus was inactivated using a replacement vector containing a UBP34truncated genomic fragment (85%of the ORF has been removed and replaced by the NptII cassette as a selectable marker).In our study,speciﬁc gene replacement at the UBP34locus occured in30%of in-tegrative transformants.The insertion of such a replacement vector in the genome results in a null mutation.UBP34-knockout plants cultivated on various media,in response to cytokinin and in stress conditions do not dis-play any phenotypic modiﬁcations.Even if old publications described interaction of mosses with mycorrhizal fungus and soil organisms as quoted in[16],little is known about the responses of P.patens to pathogenic microorganisms. Tomato spotted wilt virus has been reported to infect wild type P.patens(Kellmann J.,16.Tagung Molekularbiologie der Pﬂanzen,Dabringhause2002),but viruses able to in-fect speciﬁcally bryophytes have not been described so far [17].Biological assays based on moss behavior in response to higher plant pathogens or elicitors would be very use-ful to go further in the phenotypic characterization of the UBP34-knockout lines.However,biochemical and molecular analysis of the UBP34-knockout lines are indicative of a possible com-pensation mechanism.A new protein appears speciﬁcally photoafﬁnity labeled in the UBP34-knockout lines.This protein could be puriﬁed on NH2CPPU–Sepharose6B afﬁnity column from soluble proteins of knockout lines, whereas it was not previously puriﬁed from wild type strain.Massﬁnger printing allows the identiﬁcation of this protein as the UBP34-like protein.In addition,North-ern blot analysis indicates that the UBP34-like transcript is as abundant in wild type strain as in knockout lines. If the frequency of ESTs in cDNA library are generally consistent with mRNA abundance and can constitute an indicator of gene expression,then the expression pattern of the two genes,deduced from the NIBB-PHYSCObase, may be under developmental control.Numerous EST se-quences for both UBP34and UBP34-like are listed in the P.patens databases and although UBP34EST are present in the three cDNA libraries(auxin-and cytokinin-treated gametophytes and gametophytes that were grown without exogenous plant hormones)the UBP34-like transcript is not found in the auxin treated library and over-represented in the cytokinin-treated library[18].This developmental control may be affected in an UBP34–KO genetic back-ground.At the protein level competition for CPPU between UBP34and UBP34-like cannot account for the impossibil-ity to photolabel and purify UBP34-like in wild type strain. CPPU in photolabeling experiment is indeed in large ex-cess;moreover a photolabeling experiment with a mixture of proteins from wild type and knockout line,resulted in photolabeling of both UBP34and UBP34-like(data not shown).Therefore,photoafﬁnity labeling and afﬁnity chro-matography in knockout lines indicates that UBP34-like is either less abundant in wild type strains than in knockout lines or may be present in wild type under a non functional form at least considering CPPU binding properties.Quan-tiﬁcation of UBP34-like in both genetic backgrounds,with speciﬁc antibodies,could indicate whether this protein is affected in its translation/stability.CPPU binding proper-ties to UBP34-like could also be examined.However,the deﬁnitive role of these IPR proteins in moss could be es-tablished only when a speciﬁc functional assay is available. Host pathogen in moss is still a black box and increased knowledge in thisﬁeld will enable use of knockout facili-ties in P.patens for functional analysis of this class of IPR protein.AcknowledgementsWe thank Dr.F.Nogué,Dr.H.Barbier-Brygoo and Dr.H.Höfte for stimulating and helpful discussions.We thank。

A Fast and Accurate Plane Detection Algorithm for Large Noisy Point Clouds Using Filtered Normals

A Fast and Accurate Plane Detection Algorithm for Large Noisy Point CloudsUsing Filtered Normals and Voxel GrowingJean-Emmanuel DeschaudFranc¸ois GouletteMines ParisTech,CAOR-Centre de Robotique,Math´e matiques et Syst`e mes60Boulevard Saint-Michel75272Paris Cedex06jean-emmanuel.deschaud@mines-paristech.fr francois.goulette@mines-paristech.frAbstractWith the improvement of3D scanners,we produce point clouds with more and more points often exceeding millions of points.Then we need a fast and accurate plane detection algorithm to reduce data size.In this article,we present a fast and accurate algorithm to detect planes in unorganized point clouds usingﬁltered normals and voxel growing.Our work is based on aﬁrst step in estimating better normals at the data points,even in the presence of noise.In a second step,we compute a score of local plane in each point.Then, we select the best local seed plane and in a third step start a fast and robust region growing by voxels we call voxel growing.We have evaluated and tested our algorithm on different kinds of point cloud and compared its performance to other algorithms.1.IntroductionWith the growing availability of3D scanners,we are now able to produce large datasets with millions of points.It is necessary to reduce data size,to decrease the noise and at same time to increase the quality of the model.It is in-teresting to model planar regions of these point clouds by planes.In fact,plane detection is generally aﬁrst step of segmentation but it can be used for many applications.It is useful in computer graphics to model the environnement with basic geometry.It is used for example in modeling to detect building facades before classiﬁcation.Robots do Si-multaneous Localization and Mapping(SLAM)by detect-ing planes of the environment.In our laboratory,we wanted to detect small and large building planes in point clouds of urban environments with millions of points for modeling. As mentioned in[6],the accuracy of the plane detection is important for after-steps of the modeling pipeline.We also want to be fast to be able to process point clouds with mil-lions of points.We present a novel algorithm based on re-gion growing with improvements in normal estimation and growing process.For our method,we are generic to work on different kinds of data like point clouds fromﬁxed scan-ner or from Mobile Mapping Systems(MMS).We also aim at detecting building facades in urban point clouds or little planes like doors,even in very large data sets.Our input is an unorganized noisy point cloud and with only three”in-tuitive”parameters,we generate a set of connected compo-nents of planar regions.We evaluate our method as well as explain and analyse the signiﬁcance of each parameter. 2.Previous WorksAlthough there are many methods of segmentation in range images like in[10]or in[3],three have been thor-oughly studied for3D point clouds:region-growing, hough-transform from[14]and Random Sample Consen-sus(RANSAC)from[9].The application of recognising structures in urban laser point clouds is frequent in literature.Bauer in[4]and Boulaassal in[5]detect facades in dense3D point cloud by a RANSAC algorithm.V osselman in[23]reviews sur-face growing and3D hough transform techniques to de-tect geometric shapes.Tarsh-Kurdi in[22]detect roof planes in3D building point cloud by comparing results on hough-transform and RANSAC algorithm.They found that RANSAC is more efﬁcient than theﬁrst one.Chao Chen in[6]and Yu in[25]present algorithms of segmentation in range images for the same application of detecting planar regions in an urban scene.The method in[6]is based on a region growing algorithm in range images and merges re-sults in one labelled3D point cloud.[25]uses a method different from the three we have cited:they extract a hi-erarchical subdivision of the input image built like a graph where leaf nodes represent planar regions.There are also other methods like bayesian techniques. In[16]and[8],they obtain smoothed surface from noisy point clouds with objects modeled by probability distribu-tions and it seems possible to extend this idea to point cloud segmentation.But techniques based on bayesian statistics need to optimize global statistical model and then it is difﬁ-cult to process points cloud larger than one million points.We present below an analysis of the two main methods used in literature:RANSAC and region-growing.Hough-transform algorithm is too time consuming for our applica-tion.To compare the complexity of the algorithm,we take a point cloud of size N with only one plane P of size n.We suppose that we want to detect this plane P and we deﬁne n min the minimum size of the plane we want to detect.The size of a plane is the area of the plane.If the data density is uniform in the point cloud then the size of a plane can be speciﬁed by its number of points.2.1.RANSACRANSAC is an algorithm initially developped by Fis-chler and Bolles in[9]that allows theﬁtting of models with-out trying all possibilities.RANSAC is based on the prob-ability to detect a model using the minimal set required to estimate the model.To detect a plane with RANSAC,we choose3random points(enough to estimate a plane).We compute the plane parameters with these3points.Then a score function is used to determine how the model is good for the remaining ually,the score is the number of points belonging to the plane.With noise,a point belongs to a plane if the distance from the point to the plane is less than a parameter γ.In the end,we keep the plane with the best score.Theprobability of getting the plane in theﬁrst trial is p=(nN )3.Therefore the probability to get it in T trials is p=1−(1−(nN )3)ing equation1and supposing n minN1,we know the number T min of minimal trials to have a probability p t to get planes of size at least n min:T min=log(1−p t)log(1−(n minN))≈log(11−p t)(Nn min)3.(1)For each trial,we test all data points to compute the score of a plane.The RANSAC algorithm complexity lies inO(N(Nn min )3)when n minN1and T min→0whenn min→N.Then RANSAC is very efﬁcient in detecting large planes in noisy point clouds i.e.when the ratio n minN is 1but very slow to detect small planes in large pointclouds i.e.when n minN 1.After selecting the best model,another step is to extract the largest connected component of each plane.Connnected components mean that the min-imum distance between each point of the plane and others points is smaller(for distance)than aﬁxed parameter.Schnabel et al.[20]bring two optimizations to RANSAC:the points selection is done locally and the score function has been improved.An octree isﬁrst created from point cloud.Points used to estimate plane parameters are chosen locally at a random depth of the octree.The score function is also different from RANSAC:instead of testing all points for one model,they test only a random subset and ﬁnd the score by interpolation.The algorithm complexity lies in O(Nr4Ndn min)where r is the number of random subsets for the score function and d is the maximum octree depth. Their algorithm improves the planes detection speed but its complexity lies in O(N2)and it becomes slow on large data sets.And again we have to extract the largest connected component of each plane.2.2.Region GrowingRegion Growing algorithms work well in range images like in[18].The principle of region growing is to start with a seed region and to grow it by neighborhood when the neighbors satisfy some conditions.In range images,we have the neighbors of each point with pixel coordinates.In case of unorganized3D data,there is no information about the neighborhood in the data structure.The most common method to compute neighbors in3D is to compute a Kd-tree to search k nearest neighbors.The creation of a Kd-tree lies in O(NlogN)and the search of k nearest neighbors of one point lies in O(logN).The advantage of these region growing methods is that they are fast when there are many planes to extract,robust to noise and extract the largest con-nected component immediately.But they only use the dis-tance from point to plane to extract planes and like we will see later,it is not accurate enough to detect correct planar regions.Rabbani et al.[19]developped a method of smooth area detection that can be used for plane detection.Theyﬁrst estimate the normal of each point like in[13].The point with the minimum residual starts the region growing.They test k nearest neighbors of the last point added:if the an-gle between the normal of the point and the current normal of the plane is smaller than a parameterαthen they add this point to the smooth region.With Kd-tree for k nearest neighbors,the algorithm complexity is in O(N+nlogN). The complexity seems to be low but in worst case,when nN1,example for facade detection in point clouds,the complexity becomes O(NlogN).3.Voxel Growing3.1.OverviewIn this article,we present a new algorithm adapted to large data sets of unorganized3D points and optimized to be accurate and fast.Our plane detection method works in three steps.In theﬁrst part,we compute a better esti-mation of the normal in each point by aﬁltered weighted planeﬁtting.In a second step,we compute the score of lo-cal planarity in each point.We select the best seed point that represents a good seed plane and in the third part,we grow this seed plane by adding all points close to the plane.Thegrowing step is based on a voxel growing algorithm.The ﬁltered normals,the score function and the voxel growing are innovative contributions of our method.As an input,we need dense point clouds related to the level of detail we want to detect.As an output,we produce connected components of planes in the point cloud.This notion of connected components is linked to the data den-sity.With our method,the connected components of planes detected are linked to the parameter d of the voxel grid.Our method has 3”intuitive”parameters :d ,area min and γ.”intuitive”because there are linked to physical mea-surements.d is the voxel size used in voxel growing and also represents the connectivity of points in detected planes.γis the maximum distance between the point of a plane and the plane model,represents the plane thickness and is linked to the point cloud noise.area min represents the minimum area of planes we want to keep.3.2.Details3.2.1Local Density of Point CloudsIn a ﬁrst step,we compute the local density of point clouds like in [17].For that,we ﬁnd the radius r i of the sphere containing the k nearest neighbors of point i .Then we cal-culate ρi =kπr 2i.In our experiments,we ﬁnd that k =50is a good number of neighbors.It is important to know the lo-cal density because many laser point clouds are made with a ﬁxed resolution angle scanner and are therefore not evenly distributed.We use the local density in section 3.2.3for the score calculation.3.2.2Filtered Normal EstimationNormal estimation is an important part of our algorithm.The paper [7]presents and compares three normal estima-tion methods.They conclude that the weighted plane ﬁt-ting or WPF is the fastest and the most accurate for large point clouds.WPF is an idea of Pauly and al.in [17]that the ﬁtting plane of a point p must take into consider-ation the nearby points more than other distant ones.The normal least square is explained in [21]and is the mini-mum of ki =1(n p ·p i +d )2.The WPF is the minimum of ki =1ωi (n p ·p i +d )2where ωi =θ( p i −p )and θ(r )=e −2r 2r2i .For solving n p ,we compute the eigenvec-tor corresponding to the smallest eigenvalue of the weightedcovariance matrix C w = ki =1ωi t (p i −b w )(p i −b w )where b w is the weighted barycenter.For the three methods ex-plained in [7],we get a good approximation of normals in smooth area but we have errors in sharp corners.In ﬁg-ure 1,we have tested the weighted normal estimation on two planes with uniform noise and forming an angle of 90˚.We can see that the normal is not correct on the corners of the planes and in the red circle.To improve the normal calculation,that improves the plane detection especially on borders of planes,we propose a ﬁltering process in two phases.In a ﬁrst step,we com-pute the weighted normals (WPF)of each point like we de-scribed it above by minimizing ki =1ωi (n p ·p i +d )2.In a second step,we compute the ﬁltered normal by us-ing an adaptive local neighborhood.We compute the new weighted normal with the same sum minimization but keep-ing only points of the neighborhood whose normals from the ﬁrst step satisfy |n p ·n i |>cos (α).With this ﬁltering step,we have the same results in smooth areas and better results in sharp corners.We called our normal estimation ﬁltered weighted plane ﬁtting(FWPF).Figure 1.Weighted normal estimation of two planes with uniform noise and with 90˚angle between them.We have tested our normal estimation by computing nor-mals on synthetic data with two planes and different angles between them and with different values of the parameter α.We can see in ﬁgure 2the mean error on normal estimation for WPF and FWPF with α=20˚,30˚,40˚and 90˚.Us-ing α=90˚is the same as not doing the ﬁltering step.We see on Figure 2that α=20˚gives smaller error in normal estimation when angles between planes is smaller than 60˚and α=30˚gives best results when angle between planes is greater than 60˚.We have considered the value α=30˚as the best results because it gives the smaller mean error in normal estimation when angle between planes vary from 20˚to 90˚.Figure 3shows the normals of the planes with 90˚angle and better results in the red circle (normals are 90˚with the plane).3.2.3The score of local planarityIn many region growing algorithms,the criteria used for the score of the local ﬁtting plane is the residual,like in [18]or [19],i.e.the sum of the square of distance from points to the plane.We have a different score function to estimate local planarity.For that,we ﬁrst compute the neighbors N i of a point p with points i whose normals n i are close toFigure parison of mean error in normal estimation of two planes with α=20˚,30˚,40˚and 90˚(=Noﬁltering).Figure 3.Filtered Weighted normal estimation of two planes with uniform noise and with 90˚angle between them (α=30˚).the normal n p .More precisely,we compute N i ={p in k neighbors of i/|n i ·n p |>cos (α)}.It is a way to keep only the points which are probably on the local plane before the least square ﬁtting.Then,we compute the local plane ﬁtting of point p with N i neighbors by least squares like in [21].The set N i is a subset of N i of points belonging to the plane,i.e.the points for which the distance to the local plane is smaller than the parameter γ(to consider the noise).The score s of the local plane is the area of the local plane,i.e.the number of points ”in”the plane divided by the localdensity ρi (seen in section 3.2.1):the score s =card (N i)ρi.We take into consideration the area of the local plane as the score function and not the number of points or the residual in order to be more robust to the sampling distribution.3.2.4Voxel decompositionWe use a data structure that is the core of our region growing method.It is a voxel grid that speeds up the plane detection process.V oxels are small cubes of length d that partition the point cloud space.Every point of data belongs to a voxel and a voxel contains a list of points.We use the Octree Class Template in [2]to compute an Octree of the point cloud.The leaf nodes of the graph built are voxels of size d .Once the voxel grid has been computed,we start the plane detection algorithm.3.2.5Voxel GrowingWith the estimator of local planarity,we take the point p with the best score,i.e.the point with the maximum area of local plane.We have the model parameters of this best seed plane and we start with an empty set E of points belonging to the plane.The initial point p is in a voxel v 0.All the points in the initial voxel v 0for which the distance from the seed plane is less than γare added to the set E .Then,we compute new plane parameters by least square reﬁtting with set E .Instead of growing with k nearest neighbors,we grow with voxels.Hence we test points in 26voxel neigh-bors.This is a way to search the neighborhood in con-stant time instead of O (logN )for each neighbor like with Kd-tree.In a neighbor voxel,we add to E the points for which the distance to the current plane is smaller than γand the angle between the normal computed in each point and the normal of the plane is smaller than a parameter α:|cos (n p ,n P )|>cos (α)where n p is the normal of the point p and n P is the normal of the plane P .We have tested different values of αand we empirically found that 30˚is a good value for all point clouds.If we added at least one point in E for this voxel,we compute new plane parameters from E by least square ﬁtting and we test its 26voxel neigh-bors.It is important to perform plane least square ﬁtting in each voxel adding because the seed plane model is not good enough with noise to be used in all voxel growing,but only in surrounding voxels.This growing process is faster than classical region growing because we do not compute least square for each point added but only for each voxel added.The least square ﬁtting step must be computed very fast.We use the same method as explained in [18]with incre-mental update of the barycenter b and covariance matrix C like equation 2.We know with [21]that the barycen-ter b belongs to the least square plane and that the normal of the least square plane n P is the eigenvector of the smallest eigenvalue of C .b0=03x1C0=03x3.b n+1=1n+1(nb n+p n+1).C n+1=C n+nn+1t(pn+1−b n)(p n+1−b n).(2)where C n is the covariance matrix of a set of n points,b n is the barycenter vector of a set of n points and p n+1is the (n+1)point vector added to the set.This voxel growing method leads to a connected com-ponent set E because the points have been added by con-nected voxels.In our case,the minimum distance between one point and E is less than parameter d of our voxel grid. That is why the parameter d also represents the connectivity of points in detected planes.3.2.6Plane DetectionTo get all planes with an area of at least area min in the point cloud,we repeat these steps(best local seed plane choice and voxel growing)with all points by descending order of their score.Once we have a set E,whose area is bigger than area min,we keep it and classify all points in E.4.Results and Discussion4.1.Benchmark analysisTo test the improvements of our method,we have em-ployed the comparative framework of[12]based on range images.For that,we have converted all images into3D point clouds.All Point Clouds created have260k points. After our segmentation,we project labelled points on a seg-mented image and compare with the ground truth image. We have chosen our three parameters d,area min andγby optimizing the result of the10perceptron training image segmentation(the perceptron is portable scanner that pro-duces a range image of its environment).Bests results have been obtained with area min=200,γ=5and d=8 (units are not provided in the benchmark).We show the re-sults of the30perceptron images segmentation in table1. GT Regions are the mean number of ground truth planes over the30ground truth range images.Correct detection, over-segmentation,under-segmentation,missed and noise are the mean number of correct,over,under,missed and noised planes detected by methods.The tolerance80%is the minimum percentage of points we must have detected comparing to the ground truth to have a correct detection. More details are in[12].UE is a method from[12],UFPR is a method from[10]. It is important to notice that UE and UFPR are range image methods and our method is not well suited for range images but3D Point Cloud.Nevertheless,it is a good benchmark for comparison and we see in table1that the accuracy of our method is very close to the state of the art in range image segmentation.To evaluate the different improvements of our algorithm, we have tested different variants of our method.We have tested our method without normals(only with distance from points to plane),without voxel growing(with a classical region growing by k neighbors),without our FWPF nor-mal estimation(with WPF normal estimation),without our score function(with residual score function).The compari-son is visible on table2.We can see the difference of time computing between region growing and voxel growing.We have tested our algorithm with and without normals and we found that the accuracy cannot be achieved whithout normal computation.There is also a big difference in the correct de-tection between WPF and our FWPF normal estimation as we can see in theﬁgure4.Our FWPF normal brings a real improvement in border estimation of planes.Black points in theﬁgure are non classiﬁedpoints.Figure5.Correct Detection of our segmentation algorithm when the voxel size d changes.We would like to discuss the inﬂuence of parameters on our algorithm.We have three parameters:area min,which represents the minimum area of the plane we want to keep,γ,which represents the thickness of the plane(it is gener-aly closely tied to the noise in the point cloud and espe-cially the standard deviationσof the noise)and d,which is the minimum distance from a point to the rest of the plane. These three parameters depend on the point cloud features and the desired segmentation.For example,if we have a lot of noise,we must choose a highγvalue.If we want to detect only large planes,we set a large area min value.We also focus our analysis on the robustess of the voxel size d in our algorithm,i.e.the ratio of points vs voxels.We can see inﬁgure5the variation of the correct detection when we change the value of d.The method seems to be robust when d is between4and10but the quality decreases when d is over10.It is due to the fact that for a large voxel size d,some planes from different objects are merged into one plane.GT Regions Correct Over-Under-Missed Noise Duration(in s)detection segmentation segmentationUE14.610.00.20.3 3.8 2.1-UFPR14.611.00.30.1 3.0 2.5-Our method14.610.90.20.1 3.30.7308Table1.Average results of different segmenters at80%compare tolerance.GT Regions Correct Over-Under-Missed Noise Duration(in s) Our method detection segmentation segmentationwithout normals14.6 5.670.10.19.4 6.570 without voxel growing14.610.70.20.1 3.40.8605 without FWPF14.69.30.20.1 5.0 1.9195 without our score function14.610.30.20.1 3.9 1.2308 with all improvements14.610.90.20.1 3.30.7308 Table2.Average results of variants of our segmenter at80%compare tolerance.4.1.1Large scale dataWe have tested our method on different kinds of data.We have segmented urban data inﬁgure6from our Mobile Mapping System(MMS)described in[11].The mobile sys-tem generates10k pts/s with a density of50pts/m2and very noisy data(σ=0.3m).For this point cloud,we want to de-tect building facades.We have chosen area min=10m2, d=1m to have large connected components andγ=0.3m to cope with the noise.We have tested our method on point cloud from the Trim-ble VX scanner inﬁgure7.It is a point cloud of size40k points with only20pts/m2with less noise because it is a ﬁxed scanner(σ=0.2m).In that case,we also wanted to detect building facades and keep the same parameters ex-ceptγ=0.2m because we had less noise.We see inﬁg-ure7that we have detected two facades.By setting a larger voxel size d value like d=10m,we detect only one plane. We choose d like area min andγaccording to the desired segmentation and to the level of detail we want to extract from the point cloud.We also tested our algorithm on the point cloud from the LEICA Cyrax scanner inﬁgure8.This point cloud has been taken from AIM@SHAPE repository[1].It is a very dense point cloud from multipleﬁxed position of scanner with about400pts/m2and very little noise(σ=0.02m). In this case,we wanted to detect all the little planes to model the church in planar regions.That is why we have chosen d=0.2m,area min=1m2andγ=0.02m.Inﬁgures6,7and8,we have,on the left,input point cloud and on the right,we only keep points detected in a plane(planes are in random colors).The red points in theseﬁgures are seed plane points.We can see in theseﬁg-ures that planes are very well detected even with high noise. Table3show the information on point clouds,results with number of planes detected and duration of the algorithm.The time includes the computation of the FWPF normalsof the point cloud.We can see in table3that our algo-rithm performs linearly in time with respect to the numberof points.The choice of parameters will have little inﬂuence on time computing.The computation time is about one mil-lisecond per point whatever the size of the point cloud(we used a PC with QuadCore Q9300and2Go of RAM).The algorithm has been implented using only one thread andin-core processing.Our goal is to compare the improve-ment of plane detection between classical region growing and our region growing with better normals for more ac-curate planes and voxel growing for faster detection.Our method seems to be compatible with out-of-core implemen-tation like described in[24]or in[15].MMS Street VX Street Church Size(points)398k42k7.6MMean Density50pts/m220pts/m2400pts/m2 Number of Planes202142Total Duration452s33s6900sTime/point 1ms 1ms 1msTable3.Results on different data.5.ConclusionIn this article,we have proposed a new method of plane detection that is fast and accurate even in presence of noise. We demonstrate its efﬁciency with different kinds of data and its speed in large data sets with millions of points.Our voxel growing method has a complexity of O(N)and it is able to detect large and small planes in very large data sets and can extract them directly in connected components.Figure 4.Ground truth,Our Segmentation without and with ﬁlterednormals.Figure 6.Planes detection in street point cloud generated by MMS (d =1m,area min =10m 2,γ=0.3m ).References[1]Aim@shape repository /.6[2]Octree class template /code/octree.html.4[3] A.Bab-Hadiashar and N.Gheissari.Range image segmen-tation using surface selection criterion.2006.IEEE Trans-actions on Image Processing.1[4]J.Bauer,K.Karner,K.Schindler,A.Klaus,and C.Zach.Segmentation of building models from dense 3d point-clouds.2003.Workshop of the Austrian Association for Pattern Recognition.1[5]H.Boulaassal,ndes,P.Grussenmeyer,and F.Tarsha-Kurdi.Automatic segmentation of building facades using terrestrial laser data.2007.ISPRS Workshop on Laser Scan-ning.1[6] C.C.Chen and I.Stamos.Range image segmentationfor modeling and object detection in urban scenes.2007.3DIM2007.1[7]T.K.Dey,G.Li,and J.Sun.Normal estimation for pointclouds:A comparison study for a voronoi based method.2005.Eurographics on Symposium on Point-Based Graph-ics.3[8]J.R.Diebel,S.Thrun,and M.Brunig.A bayesian methodfor probable surface reconstruction and decimation.2006.ACM Transactions on Graphics (TOG).1[9]M.A.Fischler and R.C.Bolles.Random sample consen-sus:A paradigm for model ﬁtting with applications to image analysis and automated munications of the ACM.1,2[10]P.F.U.Gotardo,O.R.P.Bellon,and L.Silva.Range imagesegmentation by surface extraction using an improved robust estimator.2003.Proceedings of Computer Vision and Pat-tern Recognition.1,5[11] F.Goulette,F.Nashashibi,I.Abuhadrous,S.Ammoun,andurgeau.An integrated on-board laser range sensing sys-tem for on-the-way city and road modelling.2007.Interna-tional Archives of the Photogrammetry,Remote Sensing and Spacial Information Sciences.6[12] A.Hoover,G.Jean-Baptiste,and al.An experimental com-parison of range image segmentation algorithms.1996.IEEE Transactions on Pattern Analysis and Machine Intelligence.5[13]H.Hoppe,T.DeRose,T.Duchamp,J.McDonald,andW.Stuetzle.Surface reconstruction from unorganized points.1992.International Conference on Computer Graphics and Interactive Techniques.2[14]P.Hough.Method and means for recognizing complex pat-terns.1962.In US Patent.1[15]M.Isenburg,P.Lindstrom,S.Gumhold,and J.Snoeyink.Large mesh simpliﬁcation using processing sequences.2003.。

基于深度学习的松材线虫病树检测算法研究

47基于深度学习的松材线虫病树检测算法研究叶莎1，丁峰2，彭宜生1（1.三峡大学计算机与信息学院，湖北宜昌443002；2.宜昌三峡大老岭自然保护区管理局，湖北宜昌443000；）摘要：松材线虫病是一种毁灭性森林病害，具有极强的传染性，松材线虫病树的有效防治对森林资源保护具有重要意义。

采用无人机获取森林中松树的航拍正射影像，同时基于深度学习算法对松材线虫病树进行检测是对松材线虫病进行监测、治理最为有效的途径。

文章基于深度学习技术，使用不同的目标检测算法在松材线虫病树识别上的应用进行总结，展示采用不同的目标检测算法在松材线虫病树识别上取得的效果，并分析总结了不同的目标检测算法在松材线虫病树上应用存在的问题。

关键词：松材线虫病；森林资源保护；深度学习技术；目标检测中图分类号：TP309文献标识码：A 文章编号：2096-9759（2023）03-0047-03Research on detection algorithm of pine wilt disease tree based on depth learningYE Sha 1，DING Feng 2，PENG Yisheng 1（1.College of Computer and Information Technology ，China Three Gorges University,Yichang Hubei 443002；2.Yichang Three Gorges Dalaoling Nature Reserve Administration,Yichang,Hubei 443000）Abstract:Pine wilt disease is a destructive forest disease with strong infectivity.Effective control of pine wood nematode dis-ease trees is of great significance to the protection of forest resources.The most effective way to monitor and control pine wood nematode disease is to use UA V to obtain aerial orthophoto images of pine trees in the forest and detect pine wood nematode disease trees based on depth learning algorithm.Based on the deep learning technology,this paper summarizes the application of different target detection algorithms in the identification of pine wood nematode diseased trees,shows the effects of different target detection algorithms in the identification of pine wood nematode diseased trees,and analyzes and summarizes the prob-lems of different target detection algorithms in the application of pine wood nematode diseased trees.Keywords:pine wilt disease;protection of forest resources;deep learning technology;object detection0引言森林资源是我国的重要资源，松材线虫病是一种由松材线虫所引发的松树传染病，它具有传染速度快、防治困难、破坏性强的特性，通常松树染病后最快40天即可造成整株松树枯死，3-5年即可摧毁整片森林，对我国生态环境和经济造成了严重损失[1]。

人工神经网络算法基础精讲ppt课件

30
2.3学习规则
学习规则
在神经网络的学习中，各神经元的连接权值需按一定的规则
调整，这种权值调整规则称为学习规则。下面介绍几种常见的学习
规则。
1.Hebb学习规则
2.Delta（δ）学习规则
3.LMS学习规则
4.胜者为王学习规则
5.Kohonen学习规则
6.概率式学习规则
2.3学习规则
1.Hebb学习规则
突触结构示意图
1.3生物神经元的信息处理机理
电脉冲
输入
树突
细胞体形成轴突
突
输
触
出
信息处理
传输
图 12.2 生物神经元功能模型
神经元的兴奋与抑制
当传入神经元冲动，经整和使细胞膜电位升高，超过动作电位的阈值时，为兴奋状态，产生神经冲动，由轴突经神经末稍传出。当传入神经元的冲动，经整和，使细胞膜电位降低，低于阈值时，为抑制状态，不产生神经冲动。
④神经元的输出和响应是个输入值的综合作用的结果。
⑤兴奋和抑制状态，当细胞膜电位升高超过阈值时，细胞进入兴奋状态，产生神经冲动；当膜电位低于阈值时，细胞进入抑制状态。
13
1.6激活函数
神经元的描述有多种，其区别在于采用了不同的激活函数，不同的激活函数决定神经元的不同输出特性，常用的激活函数有如下几种类型：
1957年，F.Rosenblatt提出“感知器”(Perceptron)模型，第一次把神经网络的研究从纯理论的探讨付诸工程实践，掀起了人工神经网络研究的第一次高潮。
4
1.1人工神经网络发展简史
20世纪60年代以后，数字计算机的发展达到全盛时期，人们误以为数字计算机可以解决人工智能、专家系统、模式识别问题，而放松了对“感知器”的研究。于是，从20世纪60年代末期起，人工神经网络的研究进入了低潮。

T.W. ANDERSON (1971). The Statistical Analysis of Time Series. Series in Probability and Ma

425 BibliographyH.A KAIKE(1974).Markovian representation of stochastic processes and its application to the analysis of autoregressive moving average processes.Annals Institute Statistical Mathematics,vol.26,pp.363-387. B.D.O.A NDERSON and J.B.M OORE(1979).Optimal rmation and System Sciences Series, Prentice Hall,Englewood Cliffs,NJ.T.W.A NDERSON(1971).The Statistical Analysis of Time Series.Series in Probability and Mathematical Statistics,Wiley,New York.R.A NDRE-O BRECHT(1988).A new statistical approach for the automatic segmentation of continuous speech signals.IEEE Trans.Acoustics,Speech,Signal Processing,vol.ASSP-36,no1,pp.29-40.R.A NDRE-O BRECHT(1990).Reconnaissance automatique de parole`a partir de segments acoustiques et de mod`e les de Markov cach´e s.Proc.Journ´e es Etude de la Parole,Montr´e al,May1990(in French).R.A NDRE-O BRECHT and H.Y.S U(1988).Three acoustic labellings for phoneme based continuous speech recognition.Proc.Speech’88,Edinburgh,UK,pp.943-950.U.A PPEL and A.VON B RANDT(1983).Adaptive sequential segmentation of piecewise stationary time rmation Sciences,vol.29,no1,pp.27-56.L.A.A ROIAN and H.L EVENE(1950).The effectiveness of quality control procedures.Jal American Statis-tical Association,vol.45,pp.520-529.K.J.A STR¨OM and B.W ITTENMARK(1984).Computer Controlled Systems:Theory and rma-tion and System Sciences Series,Prentice Hall,Englewood Cliffs,NJ.M.B AGSHAW and R.A.J OHNSON(1975a).The effect of serial correlation on the performance of CUSUM tests-Part II.Technometrics,vol.17,no1,pp.73-80.M.B AGSHAW and R.A.J OHNSON(1975b).The inﬂuence of reference values and estimated variance on the ARL of CUSUM tests.Jal Royal Statistical Society,vol.37(B),no3,pp.413-420.M.B AGSHAW and R.A.J OHNSON(1977).Sequential procedures for detecting parameter changes in a time-series model.Jal American Statistical Association,vol.72,no359,pp.593-597.R.K.B ANSAL and P.P APANTONI-K AZAKOS(1986).An algorithm for detecting a change in a stochastic process.IEEE rmation Theory,vol.IT-32,no2,pp.227-235.G.A.B ARNARD(1959).Control charts and stochastic processes.Jal Royal Statistical Society,vol.B.21, pp.239-271.A.E.B ASHARINOV andB.S.F LEISHMAN(1962).Methods of the statistical sequential analysis and their radiotechnical applications.Sovetskoe Radio,Moscow(in Russian).M.B ASSEVILLE(1978).D´e viations par rapport au maximum:formules d’arrˆe t et martingales associ´e es. Compte-rendus du S´e minaire de Probabilit´e s,Universit´e de Rennes I.M.B ASSEVILLE(1981).Edge detection using sequential methods for change in level-Part II:Sequential detection of change in mean.IEEE Trans.Acoustics,Speech,Signal Processing,vol.ASSP-29,no1,pp.32-50.426B IBLIOGRAPHY M.B ASSEVILLE(1982).A survey of statistical failure detection techniques.In Contribution`a la D´e tectionS´e quentielle de Ruptures de Mod`e les Statistiques,Th`e se d’Etat,Universit´e de Rennes I,France(in English). M.B ASSEVILLE(1986).The two-models approach for the on-line detection of changes in AR processes. In Detection of Abrupt Changes in Signals and Dynamical Systems(M.Basseville,A.Benveniste,eds.). Lecture Notes in Control and Information Sciences,LNCIS77,Springer,New York,pp.169-215.M.B ASSEVILLE(1988).Detecting changes in signals and systems-A survey.Automatica,vol.24,pp.309-326.M.B ASSEVILLE(1989).Distance measures for signal processing and pattern recognition.Signal Process-ing,vol.18,pp.349-369.M.B ASSEVILLE and A.B ENVENISTE(1983a).Design and comparative study of some sequential jump detection algorithms for digital signals.IEEE Trans.Acoustics,Speech,Signal Processing,vol.ASSP-31, no3,pp.521-535.M.B ASSEVILLE and A.B ENVENISTE(1983b).Sequential detection of abrupt changes in spectral charac-teristics of digital signals.IEEE rmation Theory,vol.IT-29,no5,pp.709-724.M.B ASSEVILLE and A.B ENVENISTE,eds.(1986).Detection of Abrupt Changes in Signals and Dynamical Systems.Lecture Notes in Control and Information Sciences,LNCIS77,Springer,New York.M.B ASSEVILLE and I.N IKIFOROV(1991).A uniﬁed framework for statistical change detection.Proc.30th IEEE Conference on Decision and Control,Brighton,UK.M.B ASSEVILLE,B.E SPIAU and J.G ASNIER(1981).Edge detection using sequential methods for change in level-Part I:A sequential edge detection algorithm.IEEE Trans.Acoustics,Speech,Signal Processing, vol.ASSP-29,no1,pp.24-31.M.B ASSEVILLE, A.B ENVENISTE and G.M OUSTAKIDES(1986).Detection and diagnosis of abrupt changes in modal characteristics of nonstationary digital signals.IEEE rmation Theory,vol.IT-32,no3,pp.412-417.M.B ASSEVILLE,A.B ENVENISTE,G.M OUSTAKIDES and A.R OUG´E E(1987a).Detection and diagnosis of changes in the eigenstructure of nonstationary multivariable systems.Automatica,vol.23,no3,pp.479-489. M.B ASSEVILLE,A.B ENVENISTE,G.M OUSTAKIDES and A.R OUG´E E(1987b).Optimal sensor location for detecting changes in dynamical behavior.IEEE Trans.Automatic Control,vol.AC-32,no12,pp.1067-1075.M.B ASSEVILLE,A.B ENVENISTE,B.G ACH-D EVAUCHELLE,M.G OURSAT,D.B ONNECASE,P.D OREY, M.P REVOSTO and M.O LAGNON(1993).Damage monitoring in vibration mechanics:issues in diagnos-tics and predictive maintenance.Mechanical Systems and Signal Processing,vol.7,no5,pp.401-423.R.V.B EARD(1971).Failure Accommodation in Linear Systems through Self-reorganization.Ph.D.Thesis, Dept.Aeronautics and Astronautics,MIT,Cambridge,MA.A.B ENVENISTE and J.J.F UCHS(1985).Single sample modal identiﬁcation of a nonstationary stochastic process.IEEE Trans.Automatic Control,vol.AC-30,no1,pp.66-74.A.B ENVENISTE,M.B ASSEVILLE and G.M OUSTAKIDES(1987).The asymptotic local approach to change detection and model validation.IEEE Trans.Automatic Control,vol.AC-32,no7,pp.583-592.A.B ENVENISTE,M.M ETIVIER and P.P RIOURET(1990).Adaptive Algorithms and Stochastic Approxima-tions.Series on Applications of Mathematics,(A.V.Balakrishnan,I.Karatzas,M.Yor,eds.).Springer,New York.A.B ENVENISTE,M.B ASSEVILLE,L.E L G HAOUI,R.N IKOUKHAH and A.S.W ILLSKY(1992).An optimum robust approach to statistical failure detection and identiﬁcation.IFAC World Conference,Sydney, July1993.B IBLIOGRAPHY427 R.H.B ERK(1973).Some asymptotic aspects of sequential analysis.Annals Statistics,vol.1,no6,pp.1126-1138.R.H.B ERK(1975).Locally most powerful sequential test.Annals Statistics,vol.3,no2,pp.373-381.P.B ILLINGSLEY(1968).Convergence of Probability Measures.Wiley,New York.A.F.B ISSELL(1969).Cusum techniques for quality control.Applied Statistics,vol.18,pp.1-30.M.E.B IVAIKOV(1991).Control of the sample size for recursive estimation of parameters subject to abrupt changes.Automation and Remote Control,no9,pp.96-103.R.E.B LAHUT(1987).Principles and Practice of Information Theory.Addison-Wesley,Reading,MA.I.F.B LAKE and W.C.L INDSEY(1973).Level-crossing problems for random processes.IEEE r-mation Theory,vol.IT-19,no3,pp.295-315.G.B ODENSTEIN and H.M.P RAETORIUS(1977).Feature extraction from the encephalogram by adaptive segmentation.Proc.IEEE,vol.65,pp.642-652.T.B OHLIN(1977).Analysis of EEG signals with changing spectra using a short word Kalman estimator. Mathematical Biosciences,vol.35,pp.221-259.W.B¨OHM and P.H ACKL(1990).Improved bounds for the average run length of control charts based on ﬁnite weighted sums.Annals Statistics,vol.18,no4,pp.1895-1899.T.B OJDECKI and J.H OSZA(1984).On a generalized disorder problem.Stochastic Processes and their Applications,vol.18,pp.349-359.L.I.B ORODKIN and V.V.M OTTL’(1976).Algorithm forﬁnding the jump times of random process equation parameters.Automation and Remote Control,vol.37,no6,Part1,pp.23-32.A.A.B OROVKOV(1984).Theory of Mathematical Statistics-Estimation and Hypotheses Testing,Naouka, Moscow(in Russian).Translated in French under the title Statistique Math´e matique-Estimation et Tests d’Hypoth`e ses,Mir,Paris,1987.G.E.P.B OX and G.M.J ENKINS(1970).Time Series Analysis,Forecasting and Control.Series in Time Series Analysis,Holden-Day,San Francisco.A.VON B RANDT(1983).Detecting and estimating parameters jumps using ladder algorithms and likelihood ratio test.Proc.ICASSP,Boston,MA,pp.1017-1020.A.VON B RANDT(1984).Modellierung von Signalen mit Sprunghaft Ver¨a nderlichem Leistungsspektrum durch Adaptive Segmentierung.Doctor-Engineer Dissertation,M¨u nchen,RFA(in German).S.B RAUN,ed.(1986).Mechanical Signature Analysis-Theory and Applications.Academic Press,London. L.B REIMAN(1968).Probability.Series in Statistics,Addison-Wesley,Reading,MA.G.S.B RITOV and L.A.M IRONOVSKI(1972).Diagnostics of linear systems of automatic regulation.Tekh. Kibernetics,vol.1,pp.76-83.B.E.B RODSKIY and B.S.D ARKHOVSKIY(1992).Nonparametric Methods in Change-point Problems. Kluwer Academic,Boston.L.D.B ROEMELING(1982).Jal Econometrics,vol.19,Special issue on structural change in Econometrics. L.D.B ROEMELING and H.T SURUMI(1987).Econometrics and Structural Change.Dekker,New York. D.B ROOK and D.A.E VANS(1972).An approach to the probability distribution of Cusum run length. Biometrika,vol.59,pp.539-550.J.B RUNET,D.J AUME,M.L ABARR`E RE,A.R AULT and M.V ERG´E(1990).D´e tection et Diagnostic de Pannes.Trait´e des Nouvelles Technologies,S´e rie Diagnostic et Maintenance,Herm`e s,Paris(in French).428B IBLIOGRAPHY S.P.B RUZZONE and M.K AVEH(1984).Information tradeoffs in using the sample autocorrelation function in ARMA parameter estimation.IEEE Trans.Acoustics,Speech,Signal Processing,vol.ASSP-32,no4, pp.701-715.A.K.C AGLAYAN(1980).Necessary and sufﬁcient conditions for detectability of jumps in linear systems. IEEE Trans.Automatic Control,vol.AC-25,no4,pp.833-834.A.K.C AGLAYAN and R.E.L ANCRAFT(1983).Reinitialization issues in fault tolerant systems.Proc.Amer-ican Control Conf.,pp.952-955.A.K.C AGLAYAN,S.M.A LLEN and K.W EHMULLER(1988).Evaluation of a second generation reconﬁgu-ration strategy for aircraftﬂight control systems subjected to actuator failure/surface damage.Proc.National Aerospace and Electronic Conference,Dayton,OH.P.E.C AINES(1988).Linear Stochastic Systems.Series in Probability and Mathematical Statistics,Wiley, New York.M.J.C HEN and J.P.N ORTON(1987).Estimation techniques for tracking rapid parameter changes.Intern. Jal Control,vol.45,no4,pp.1387-1398.W.K.C HIU(1974).The economic design of cusum charts for controlling normal mean.Applied Statistics, vol.23,no3,pp.420-433.E.Y.C HOW(1980).A Failure Detection System Design Methodology.Ph.D.Thesis,M.I.T.,L.I.D.S.,Cam-bridge,MA.E.Y.C HOW and A.S.W ILLSKY(1984).Analytical redundancy and the design of robust failure detection systems.IEEE Trans.Automatic Control,vol.AC-29,no3,pp.689-691.Y.S.C HOW,H.R OBBINS and D.S IEGMUND(1971).Great Expectations:The Theory of Optimal Stop-ping.Houghton-Mifﬂin,Boston.R.N.C LARK,D.C.F OSTH and V.M.W ALTON(1975).Detection of instrument malfunctions in control systems.IEEE Trans.Aerospace Electronic Systems,vol.AES-11,pp.465-473.A.C OHEN(1987).Biomedical Signal Processing-vol.1:Time and Frequency Domain Analysis;vol.2: Compression and Automatic Recognition.CRC Press,Boca Raton,FL.J.C ORGE and F.P UECH(1986).Analyse du rythme cardiaque foetal par des m´e thodes de d´e tection de ruptures.Proc.7th INRIA Int.Conf.Analysis and optimization of Systems.Antibes,FR(in French).D.R.C OX and D.V.H INKLEY(1986).Theoretical Statistics.Chapman and Hall,New York.D.R.C OX and H.D.M ILLER(1965).The Theory of Stochastic Processes.Wiley,New York.S.V.C ROWDER(1987).A simple method for studying run-length distributions of exponentially weighted moving average charts.Technometrics,vol.29,no4,pp.401-407.H.C S¨ORG¨O and L.H ORV´ATH(1988).Nonparametric methods for change point problems.In Handbook of Statistics(P.R.Krishnaiah,C.R.Rao,eds.),vol.7,Elsevier,New York,pp.403-425.R.B.D AVIES(1973).Asymptotic inference in stationary gaussian time series.Advances Applied Probability, vol.5,no3,pp.469-497.J.C.D ECKERT,M.N.D ESAI,J.J.D EYST and A.S.W ILLSKY(1977).F-8DFBW sensor failure identiﬁcation using analytical redundancy.IEEE Trans.Automatic Control,vol.AC-22,no5,pp.795-803.M.H.D E G ROOT(1970).Optimal Statistical Decisions.Series in Probability and Statistics,McGraw-Hill, New York.J.D ESHAYES and D.P ICARD(1979).Tests de ruptures dans un mod`e pte-Rendus de l’Acad´e mie des Sciences,vol.288,Ser.A,pp.563-566(in French).B IBLIOGRAPHY429 J.D ESHAYES and D.P ICARD(1983).Ruptures de Mod`e les en Statistique.Th`e ses d’Etat,Universit´e deParis-Sud,Orsay,France(in French).J.D ESHAYES and D.P ICARD(1986).Off-line statistical analysis of change-point models using non para-metric and likelihood methods.In Detection of Abrupt Changes in Signals and Dynamical Systems(M. Basseville,A.Benveniste,eds.).Lecture Notes in Control and Information Sciences,LNCIS77,Springer, New York,pp.103-168.B.D EVAUCHELLE-G ACH(1991).Diagnostic M´e canique des Fatigues sur les Structures Soumises`a des Vibrations en Ambiance de Travail.Th`e se de l’Universit´e Paris IX Dauphine(in French).B.D EVAUCHELLE-G ACH,M.B ASSEVILLE and A.B ENVENISTE(1991).Diagnosing mechanical changes in vibrating systems.Proc.SAFEPROCESS’91,Baden-Baden,FRG,pp.85-89.R.D I F RANCESCO(1990).Real-time speech segmentation using pitch and convexity jump models:applica-tion to variable rate speech coding.IEEE Trans.Acoustics,Speech,Signal Processing,vol.ASSP-38,no5, pp.741-748.X.D ING and P.M.F RANK(1990).Fault detection via factorization approach.Systems and Control Letters, vol.14,pp.431-436.J.L.D OOB(1953).Stochastic Processes.Wiley,New York.V.D RAGALIN(1988).Asymptotic solutions in detecting a change in distribution under an unknown param-eter.Statistical Problems of Control,Issue83,Vilnius,pp.45-52.B.D UBUISSON(1990).Diagnostic et Reconnaissance des Formes.Trait´e des Nouvelles Technologies,S´e rie Diagnostic et Maintenance,Herm`e s,Paris(in French).A.J.D UNCAN(1986).Quality Control and Industrial Statistics,5th edition.Richard D.Irwin,Inc.,Home-wood,IL.J.D URBIN(1971).Boundary-crossing probabilities for the Brownian motion and Poisson processes and techniques for computing the power of the Kolmogorov-Smirnov test.Jal Applied Probability,vol.8,pp.431-453.J.D URBIN(1985).Theﬁrst passage density of the crossing of a continuous Gaussian process to a general boundary.Jal Applied Probability,vol.22,no1,pp.99-122.A.E MAMI-N AEINI,M.M.A KHTER and S.M.R OCK(1988).Effect of model uncertainty on failure detec-tion:the threshold selector.IEEE Trans.Automatic Control,vol.AC-33,no12,pp.1106-1115.J.D.E SARY,F.P ROSCHAN and D.W.W ALKUP(1967).Association of random variables with applications. Annals Mathematical Statistics,vol.38,pp.1466-1474.W.D.E WAN and K.W.K EMP(1960).Sampling inspection of continuous processes with no autocorrelation between successive results.Biometrika,vol.47,pp.263-280.G.F AVIER and A.S MOLDERS(1984).Adaptive smoother-predictors for tracking maneuvering targets.Proc. 23rd Conf.Decision and Control,Las Vegas,NV,pp.831-836.W.F ELLER(1966).An Introduction to Probability Theory and Its Applications,vol.2.Series in Probability and Mathematical Statistics,Wiley,New York.R.A.F ISHER(1925).Theory of statistical estimation.Proc.Cambridge Philosophical Society,vol.22, pp.700-725.M.F ISHMAN(1988).Optimization of the algorithm for the detection of a disorder,based on the statistic of exponential smoothing.In Statistical Problems of Control,Issue83,Vilnius,pp.146-151.R.F LETCHER(1980).Practical Methods of Optimization,2volumes.Wiley,New York.P.M.F RANK(1990).Fault diagnosis in dynamic systems using analytical and knowledge based redundancy -A survey and new results.Automatica,vol.26,pp.459-474.430B IBLIOGRAPHY P.M.F RANK(1991).Enhancement of robustness in observer-based fault detection.Proc.SAFEPRO-CESS’91,Baden-Baden,FRG,pp.275-287.P.M.F RANK and J.W¨UNNENBERG(1989).Robust fault diagnosis using unknown input observer schemes. In Fault Diagnosis in Dynamic Systems-Theory and Application(R.Patton,P.Frank,R.Clark,eds.). International Series in Systems and Control Engineering,Prentice Hall International,London,UK,pp.47-98.K.F UKUNAGA(1990).Introduction to Statistical Pattern Recognition,2d ed.Academic Press,New York. S.I.G ASS(1958).Linear Programming:Methods and Applications.McGraw Hill,New York.W.G E and C.Z.F ANG(1989).Extended robust observation approach for failure isolation.Int.Jal Control, vol.49,no5,pp.1537-1553.W.G ERSCH(1986).Two applications of parametric time series modeling methods.In Mechanical Signature Analysis-Theory and Applications(S.Braun,ed.),chap.10.Academic Press,London.J.J.G ERTLER(1988).Survey of model-based failure detection and isolation in complex plants.IEEE Control Systems Magazine,vol.8,no6,pp.3-11.J.J.G ERTLER(1991).Analytical redundancy methods in fault detection and isolation.Proc.SAFEPRO-CESS’91,Baden-Baden,FRG,pp.9-22.B.K.G HOSH(1970).Sequential Tests of Statistical Hypotheses.Addison-Wesley,Cambridge,MA.I.N.G IBRA(1975).Recent developments in control charts techniques.Jal Quality Technology,vol.7, pp.183-192.J.P.G ILMORE and R.A.M C K ERN(1972).A redundant strapdown inertial reference unit(SIRU).Jal Space-craft,vol.9,pp.39-47.M.A.G IRSHICK and H.R UBIN(1952).A Bayes approach to a quality control model.Annals Mathematical Statistics,vol.23,pp.114-125.A.L.G OEL and S.M.W U(1971).Determination of the ARL and a contour nomogram for CUSUM charts to control normal mean.Technometrics,vol.13,no2,pp.221-230.P.L.G OLDSMITH and H.W HITFIELD(1961).Average run lengths in cumulative chart quality control schemes.Technometrics,vol.3,pp.11-20.G.C.G OODWIN and K.S.S IN(1984).Adaptive Filtering,Prediction and rmation and System Sciences Series,Prentice Hall,Englewood Cliffs,NJ.R.M.G RAY and L.D.D AVISSON(1986).Random Processes:a Mathematical Approach for Engineers. Information and System Sciences Series,Prentice Hall,Englewood Cliffs,NJ.C.G UEGUEN and L.L.S CHARF(1980).Exact maximum likelihood identiﬁcation for ARMA models:a signal processing perspective.Proc.1st EUSIPCO,Lausanne.D.E.G USTAFSON, A.S.W ILLSKY,J.Y.W ANG,M.C.L ANCASTER and J.H.T RIEBWASSER(1978). ECG/VCG rhythm diagnosis using statistical signal analysis.Part I:Identiﬁcation of persistent rhythms. Part II:Identiﬁcation of transient rhythms.IEEE Trans.Biomedical Engineering,vol.BME-25,pp.344-353 and353-361.F.G USTAFSSON(1991).Optimal segmentation of linear regression parameters.Proc.IFAC/IFORS Symp. Identiﬁcation and System Parameter Estimation,Budapest,pp.225-229.T.H¨AGGLUND(1983).New Estimation Techniques for Adaptive Control.Ph.D.Thesis,Lund Institute of Technology,Lund,Sweden.T.H¨AGGLUND(1984).Adaptive control of systems subject to large parameter changes.Proc.IFAC9th World Congress,Budapest.B IBLIOGRAPHY431 P.H ALL and C.C.H EYDE(1980).Martingale Limit Theory and its Application.Probability and Mathemat-ical Statistics,a Series of Monographs and Textbooks,Academic Press,New York.W.J.H ALL,R.A.W IJSMAN and J.K.G HOSH(1965).The relationship between sufﬁciency and invariance with applications in sequential analysis.Ann.Math.Statist.,vol.36,pp.576-614.E.J.H ANNAN and M.D EISTLER(1988).The Statistical Theory of Linear Systems.Series in Probability and Mathematical Statistics,Wiley,New York.J.D.H EALY(1987).A note on multivariate CuSum procedures.Technometrics,vol.29,pp.402-412.D.M.H IMMELBLAU(1970).Process Analysis by Statistical Methods.Wiley,New York.D.M.H IMMELBLAU(1978).Fault Detection and Diagnosis in Chemical and Petrochemical Processes. Chemical Engineering Monographs,vol.8,Elsevier,Amsterdam.W.G.S.H INES(1976a).A simple monitor of a system with sudden parameter changes.IEEE r-mation Theory,vol.IT-22,no2,pp.210-216.W.G.S.H INES(1976b).Improving a simple monitor of a system with sudden parameter changes.IEEE rmation Theory,vol.IT-22,no4,pp.496-499.D.V.H INKLEY(1969).Inference about the intersection in two-phase regression.Biometrika,vol.56,no3, pp.495-504.D.V.H INKLEY(1970).Inference about the change point in a sequence of random variables.Biometrika, vol.57,no1,pp.1-17.D.V.H INKLEY(1971).Inference about the change point from cumulative sum-tests.Biometrika,vol.58, no3,pp.509-523.D.V.H INKLEY(1971).Inference in two-phase regression.Jal American Statistical Association,vol.66, no336,pp.736-743.J.R.H UDDLE(1983).Inertial navigation system error-model considerations in Kalmanﬁltering applica-tions.In Control and Dynamic Systems(C.T.Leondes,ed.),Academic Press,New York,pp.293-339.J.S.H UNTER(1986).The exponentially weighted moving average.Jal Quality Technology,vol.18,pp.203-210.I.A.I BRAGIMOV and R.Z.K HASMINSKII(1981).Statistical Estimation-Asymptotic Theory.Applications of Mathematics Series,vol.16.Springer,New York.R.I SERMANN(1984).Process fault detection based on modeling and estimation methods-A survey.Auto-matica,vol.20,pp.387-404.N.I SHII,A.I WATA and N.S UZUMURA(1979).Segmentation of nonstationary time series.Int.Jal Systems Sciences,vol.10,pp.883-894.J.E.J ACKSON and R.A.B RADLEY(1961).Sequential and tests.Annals Mathematical Statistics, vol.32,pp.1063-1077.B.J AMES,K.L.J AMES and D.S IEGMUND(1988).Conditional boundary crossing probabilities with appli-cations to change-point problems.Annals Probability,vol.16,pp.825-839.M.K.J EERAGE(1990).Reliability analysis of fault-tolerant IMU architectures with redundant inertial sen-sors.IEEE Trans.Aerospace and Electronic Systems,vol.AES-5,no.7,pp.23-27.N.L.J OHNSON(1961).A simple theoretical approach to cumulative sum control charts.Jal American Sta-tistical Association,vol.56,pp.835-840.N.L.J OHNSON and F.C.L EONE(1962).Cumulative sum control charts:mathematical principles applied to their construction and use.Parts I,II,III.Industrial Quality Control,vol.18,pp.15-21;vol.19,pp.29-36; vol.20,pp.22-28.432B IBLIOGRAPHY R.A.J OHNSON and M.B AGSHAW(1974).The effect of serial correlation on the performance of CUSUM tests-Part I.Technometrics,vol.16,no.1,pp.103-112.H.L.J ONES(1973).Failure Detection in Linear Systems.Ph.D.Thesis,Dept.Aeronautics and Astronautics, MIT,Cambridge,MA.R.H.J ONES,D.H.C ROWELL and L.E.K APUNIAI(1970).Change detection model for serially correlated multivariate data.Biometrics,vol.26,no2,pp.269-280.M.J URGUTIS(1984).Comparison of the statistical properties of the estimates of the change times in an autoregressive process.In Statistical Problems of Control,Issue65,Vilnius,pp.234-243(in Russian).T.K AILATH(1980).Linear rmation and System Sciences Series,Prentice Hall,Englewood Cliffs,NJ.L.V.K ANTOROVICH and V.I.K RILOV(1958).Approximate Methods of Higher Analysis.Interscience,New York.S.K ARLIN and H.M.T AYLOR(1975).A First Course in Stochastic Processes,2d ed.Academic Press,New York.S.K ARLIN and H.M.T AYLOR(1981).A Second Course in Stochastic Processes.Academic Press,New York.D.K AZAKOS and P.P APANTONI-K AZAKOS(1980).Spectral distance measures between gaussian pro-cesses.IEEE Trans.Automatic Control,vol.AC-25,no5,pp.950-959.K.W.K EMP(1958).Formula for calculating the operating characteristic and average sample number of some sequential tests.Jal Royal Statistical Society,vol.B-20,no2,pp.379-386.K.W.K EMP(1961).The average run length of the cumulative sum chart when a V-mask is used.Jal Royal Statistical Society,vol.B-23,pp.149-153.K.W.K EMP(1967a).Formal expressions which can be used for the determination of operating character-istics and average sample number of a simple sequential test.Jal Royal Statistical Society,vol.B-29,no2, pp.248-262.K.W.K EMP(1967b).A simple procedure for determining upper and lower limits for the average sample run length of a cumulative sum scheme.Jal Royal Statistical Society,vol.B-29,no2,pp.263-265.D.P.K ENNEDY(1976).Some martingales related to cumulative sum tests and single server queues.Stochas-tic Processes and Appl.,vol.4,pp.261-269.T.H.K ERR(1980).Statistical analysis of two-ellipsoid overlap test for real time failure detection.IEEE Trans.Automatic Control,vol.AC-25,no4,pp.762-772.T.H.K ERR(1982).False alarm and correct detection probabilities over a time interval for restricted classes of failure detection algorithms.IEEE rmation Theory,vol.IT-24,pp.619-631.T.H.K ERR(1987).Decentralizedﬁltering and redundancy management for multisensor navigation.IEEE Trans.Aerospace and Electronic systems,vol.AES-23,pp.83-119.Minor corrections on p.412and p.599 (May and July issues,respectively).R.A.K HAN(1978).Wald’s approximations to the average run length in cusum procedures.Jal Statistical Planning and Inference,vol.2,no1,pp.63-77.R.A.K HAN(1979).Someﬁrst passage problems related to cusum procedures.Stochastic Processes and Applications,vol.9,no2,pp.207-215.R.A.K HAN(1981).A note on Page’s two-sided cumulative sum procedures.Biometrika,vol.68,no3, pp.717-719.B IBLIOGRAPHY433 V.K IREICHIKOV,V.M ANGUSHEV and I.N IKIFOROV(1990).Investigation and application of CUSUM algorithms to monitoring of sensors.In Statistical Problems of Control,Issue89,Vilnius,pp.124-130(in Russian).G.K ITAGAWA and W.G ERSCH(1985).A smoothness prior time-varying AR coefﬁcient modeling of non-stationary covariance time series.IEEE Trans.Automatic Control,vol.AC-30,no1,pp.48-56.N.K LIGIENE(1980).Probabilities of deviations of the change point estimate in statistical models.In Sta-tistical Problems of Control,Issue83,Vilnius,pp.80-86(in Russian).N.K LIGIENE and L.T ELKSNYS(1983).Methods of detecting instants of change of random process prop-erties.Automation and Remote Control,vol.44,no10,Part II,pp.1241-1283.J.K ORN,S.W.G ULLY and A.S.W ILLSKY(1982).Application of the generalized likelihood ratio algorithm to maneuver detection and estimation.Proc.American Control Conf.,Arlington,V A,pp.792-798.P.R.K RISHNAIAH and B.Q.M IAO(1988).Review about estimation of change points.In Handbook of Statistics(P.R.Krishnaiah,C.R.Rao,eds.),vol.7,Elsevier,New York,pp.375-402.P.K UDVA,N.V ISWANADHAM and A.R AMAKRISHNAN(1980).Observers for linear systems with unknown inputs.IEEE Trans.Automatic Control,vol.AC-25,no1,pp.113-115.S.K ULLBACK(1959).Information Theory and Statistics.Wiley,New York(also Dover,New York,1968). K.K UMAMARU,S.S AGARA and T.S¨ODERSTR¨OM(1989).Some statistical methods for fault diagnosis for dynamical systems.In Fault Diagnosis in Dynamic Systems-Theory and Application(R.Patton,P.Frank,R. Clark,eds.).International Series in Systems and Control Engineering,Prentice Hall International,London, UK,pp.439-476.A.K USHNIR,I.N IKIFOROV and I.S AVIN(1983).Statistical adaptive algorithms for automatic detection of seismic signals-Part I:One-dimensional case.In Earthquake Prediction and the Study of the Earth Structure,Naouka,Moscow(Computational Seismology,vol.15),pp.154-159(in Russian).L.L ADELLI(1990).Diffusion approximation for a pseudo-likelihood test process with application to de-tection of change in stochastic system.Stochastics and Stochastics Reports,vol.32,pp.1-25.T.L.L A¨I(1974).Control charts based on weighted sums.Annals Statistics,vol.2,no1,pp.134-147.T.L.L A¨I(1981).Asymptotic optimality of invariant sequential probability ratio tests.Annals Statistics, vol.9,no2,pp.318-333.D.G.L AINIOTIS(1971).Joint detection,estimation,and system identiﬁrmation and Control, vol.19,pp.75-92.M.R.L EADBETTER,G.L INDGREN and H.R OOTZEN(1983).Extremes and Related Properties of Random Sequences and Processes.Series in Statistics,Springer,New York.L.L E C AM(1960).Locally asymptotically normal families of distributions.Univ.California Publications in Statistics,vol.3,pp.37-98.L.L E C AM(1986).Asymptotic Methods in Statistical Decision Theory.Series in Statistics,Springer,New York.E.L.L EHMANN(1986).Testing Statistical Hypotheses,2d ed.Wiley,New York.J.P.L EHOCZKY(1977).Formulas for stopped diffusion processes with stopping times based on the maxi-mum.Annals Probability,vol.5,no4,pp.601-607.H.R.L ERCHE(1980).Boundary Crossing of Brownian Motion.Lecture Notes in Statistics,vol.40,Springer, New York.L.L JUNG(1987).System Identiﬁcation-Theory for the rmation and System Sciences Series, Prentice Hall,Englewood Cliffs,NJ.。

朴素贝叶斯增量学习在病毒上报分析中的应用

ｔｅｃａｓｅ．ｎｔｒｅｐｏａｉｔｆｒｈｌｓｉｃｔｎｏｒｆｒｎｉｌｅｒｅａｌｓｉｈｓｐｐｒｗｒｐｓａｖａｅｉｎｉ — ｈｌｓｉｒＩｅｍｓｏｔｒｂｂｌｙｏｉｔｃａｓｆａｉｆｅｅｅｔｌｌａｎｄｓｍｐｅ，ｎｔｉａｅｅｐｏｏｅａｎ＇ｅｂｙｓｎｉｆｆｈｉｇｉｏｐａｙ￣＇ａｃｅｎａａｎｎｇｒｈｂｓｄｏｈｒｄｂｌｙｏａｌｓｌｓｉｅｅｕｔＩｈｌｏｉｍ，ｈｓｔｉｈｒｃｅｉｉｔｎｌａｎｎｒｍｅｔｌｌｒｉｇａｏｉｍａｅｎｔｅｃｅｉｉｔｆｓｍｐｅ ’ｃａｓｄｒｓｌ．ｎｔｅａｇｒｈｔｏｅｗｉｈｇｅｒｄｂｌｙｉｅｒｉｇｅｌｔｉｉｆｔｈｉ
ｔｅａｇｒｔｍｏｕｏｔｎｌｓｓａｄｄｔｃｉｎｏｕｐｃｏｓｓｍｐｅ．ｘｅｍｅｔｅｕｔｎｉａｅａｅｃａｓｆｒｆｎｒｍｅｔｌａｎｎｈｌｏｉｈｆｒａｔｍａｉａａｙｉｎｅｅｔｆｓｓｉｉｕａｌｓＥｐｒｎｓｌｉｄｃｔｓｔｔｈｌｓｉｅｃｅｎａｙｌｒｉｇｃｏｉｒｈｔｉｏｉｌｅｔｒｕｈｔｅａｇｒｈｏｔｅｏｍｓｔｅｃａｓｅｆａｄｍｌｎｒｍｅｔｌｅｒｉｇｉｅｅｔｎｅｆｃ．ｈｏｇｈｏｉｍｕｐｒｒｈｌｓｉｒｏｎｏｙｉｃｅｎａａｎｎｎｄｔｃｉｆｔｌｔｆｉｆｒｌｏｅ

生物信息学考试试题

生物信息学考试试题一、选择题（每题 3 分，共 30 分）1、以下哪种不是常见的生物信息学数据库？（）A GenBankB SWISSPROTC PubMedD Baidu2、在 DNA 序列分析中，以下哪个不是用于序列比对的算法？（）A NeedlemanWunsch 算法B SmithWaterman 算法C BLAST 算法D Fourier 变换算法3、蛋白质结构预测的方法不包括（）A 同源建模B 从头预测C 折叠识别D 随机模拟4、以下哪种不是基因表达数据分析的常用方法？（）A 聚类分析B 主成分分析C 判别分析D 回归分析5、生物信息学中，用于预测蛋白质功能的方法有（）A 基于序列相似性B 基于结构相似性C 基于基因共表达D 以上都是6、在基因组学中，以下哪个不是测序技术？（）A Sanger 测序B 二代测序C 三代测序D 四代测序7、系统发生树构建的方法不包括（）A 距离法B 最大简约法C 最大似然法D 最小二乘法8、以下哪种不是生物信息学中常用的编程语言？（）A PythonB JavaC C+＋D Visual Basic9、以下哪个不是生物信息学在医学领域的应用？（）A 疾病诊断B 药物研发C 医疗美容D 个性化医疗10、生物信息学中，处理大规模数据常用的工具是（）A ExcelB R 语言C SPSSD Word二、填空题（每题 2 分，共 20 分）1、生物信息学是一门融合了生物学、计算机科学和（）的交叉学科。

2、常见的核酸序列格式有 FASTA 和（）。

3、蛋白质的二级结构包括α螺旋、β折叠和（）等。

4、基因芯片技术是一种（）分析技术。

5、序列比对的目的是寻找两个或多个序列之间的（）。

6、人类基因组计划的主要目标是测定人类基因组的（）序列。

7、生物信息学中的隐马尔可夫模型主要用于（）。

8、系统发生分析中，外群的作用是（）。

9、蛋白质相互作用网络分析有助于理解（）。

10、生物信息学数据库可以分为一级数据库和（）数据库。

distributed representations of words and phrases and their compositionality

Tomas MikolovGoogle Inc.Mountain View mikolov@Ilya SutskeverGoogle Inc.Mountain Viewilyasu@Kai ChenGoogle Inc.Mountain Viewkai@Greg CorradoGoogle Inc.Mountain View gcorrado@Jeffrey DeanGoogle Inc.Mountain View jeff@AbstractThe recently introduced continuous Skip-gram model is an efﬁcient method forlearning high-quality distributed vector representations that capture a large num-ber of precise syntactic and semantic word relationships.In this paper we presentseveral extensions that improve both the quality of the vectors and the trainingspeed.By subsampling of the frequent words we obtain signiﬁcant speedup andalso learn more regular word representations.We also describe a simple alterna-tive to the hierarchical softmax called negative sampling.An inherent limitation of word representations is their indifference to word orderand their inability to represent idiomatic phrases.For example,the meanings of“Canada”and“Air”cannot be easily combined to obtain“Air Canada”.Motivatedby this example,we present a simple method forﬁnding phrases in text,and showthat learning good vector representations for millions of phrases is possible.1IntroductionDistributed representations of words in a vector space help learning algorithms to achieve better performance in natural language processing tasks by grouping similar words.One of the earliest use of word representations dates back to1986due to Rumelhart,Hinton,and Williams[13].This idea has since been applied to statistical language modeling with considerable success[1].The follow up work includes applications to automatic speech recognition and machine translation[14,7],and a wide range of NLP tasks[2,20,15,3,18,19,9].Recently,Mikolov et al.[8]introduced the Skip-gram model,an efﬁcient method for learning high-quality vector representations of words from large amounts of unstructured text data.Unlike most of the previously used neural network architectures for learning word vectors,training of the Skip-gram model(see Figure1)does not involve dense matrix multiplications.This makes the training extremely efﬁcient:an optimized single-machine implementation can train on more than100billion words in one day.The word representations computed using neural networks are very interesting because the learned vectors explicitly encode many linguistic regularities and patterns.Somewhat surprisingly,many of these patterns can be represented as linear translations.For example,the result of a vector calcula-tion vec(“Madrid”)-vec(“Spain”)+vec(“France”)is closer to vec(“Paris”)than to any other word vector[9,8].Figure1:The Skip-gram vector representations that are good at predictingIn this paper we We show that sub-sampling of frequent(around2x-10x),and improves accuracy of we present a simpli-ﬁed variant of Noise model that results in faster training and better vector representations for frequent words,compared to more complex hierarchical softmax that was used in the prior work[8].Word representations are limited by their inability to represent idiomatic phrases that are not com-positions of the individual words.For example,“Boston Globe”is a newspaper,and so it is not a natural combination of the meanings of“Boston”and“Globe”.Therefore,using vectors to repre-sent the whole phrases makes the Skip-gram model considerably more expressive.Other techniques that aim to represent meaning of sentences by composing the word vectors,such as the recursive autoencoders[15],would also beneﬁt from using phrase vectors instead of the word vectors.The extension from word based to phrase based models is relatively simple.First we identify a large number of phrases using a data-driven approach,and then we treat the phrases as individual tokens during the training.To evaluate the quality of the phrase vectors,we developed a test set of analogi-cal reasoning tasks that contains both words and phrases.A typical analogy pair from our test set is “Montreal”:“Montreal Canadiens”::“Toronto”:“Toronto Maple Leafs”.It is considered to have been answered correctly if the nearest representation to vec(“Montreal Canadiens”)-vec(“Montreal”)+ vec(“Toronto”)is vec(“Toronto Maple Leafs”).Finally,we describe another interesting property of the Skip-gram model.We found that simple vector addition can often produce meaningful results.For example,vec(“Russia”)+vec(“river”)is close to vec(“V olga River”),and vec(“Germany”)+vec(“capital”)is close to vec(“Berlin”).This compositionality suggests that a non-obvious degree of language understanding can be obtained by using basic mathematical operations on the word vector representations.2The Skip-gram ModelThe training objective of the Skip-gram model is toﬁnd word representations that are useful for predicting the surrounding words in a sentence or a document.More formally,given a sequence of training words w1,w2,w3,...,w T,the objective of the Skip-gram model is to maximize the average log probability1training time.The basic Skip-gram formulation deﬁnes p(w t+j|w t)using the softmax function:exp v′w O⊤v w Ip(w O|w I)=-2-1.5-1-0.5 0 0.511.5 2-2-1.5-1-0.5 0 0.5 1 1.5 2Country and Capital Vectors Projected by PCAChinaJapanFranceRussiaGermanyItalySpainGreece TurkeyBeijingParis Tokyo PolandMoscow Portugal Berlin Rome Athens MadridAnkara Warsaw LisbonFigure 2:Two-dimensional PCA projection of the 1000-dimensional Skip-gram vectors of countries and their capital cities.The ﬁgure illustrates ability of the model to automatically organize concepts and learn implicitly the relationships between them,as during the training we did not provide any supervised information about what a capital city means.which is used to replace every log P (w O |w I )term in the Skip-gram objective.Thus the task is to distinguish the target word w O from draws from the noise distribution P n (w )using logistic regres-sion,where there are k negative samples for each data sample.Our experiments indicate that values of k in the range 5–20are useful for small training datasets,while for large datasets the k can be as small as 2–5.The main difference between the Negative sampling and NCE is that NCE needs both samples and the numerical probabilities of the noise distribution,while Negative sampling uses only samples.And while NCE approximately maximizes the log probability of the softmax,this property is not important for our application.Both NCE and NEG have the noise distribution P n (w )as a free parameter.We investigated a number of choices for P n (w )and found that the unigram distribution U (w )raised to the 3/4rd power (i.e.,U (w )3/4/Z )outperformed signiﬁcantly the unigram and the uniform distributions,for both NCE and NEG on every task we tried including language modeling (not reported here).2.3Subsampling of Frequent WordsIn very large corpora,the most frequent words can easily occur hundreds of millions of times (e.g.,“in”,“the”,and “a”).Such words usually provide less information value than the rare words.For example,while the Skip-gram model beneﬁts from observing the co-occurrences of “France”and “Paris”,it beneﬁts much less from observing the frequent co-occurrences of “France”and “the”,as nearly every word co-occurs frequently within a sentence with “the”.This idea can also be applied in the opposite direction;the vector representations of frequent words do not change signiﬁcantly after training on several million examples.To counter the imbalance between the rare and frequent words,we used a simple subsampling ap-proach:each word w i in the training set is discarded with probability computed by the formulaP (w i )=1− f (w i )(5)Method Syntactic[%]Semantic[%]NEG-563549761 HS-Huffman53403853NEG-561583661 HS-Huffman5259/p/word2vec/source/browse/trunk/questions-words.txtNewspapersNHL TeamsNBA TeamsAirlinesCompany executives.(6)count(w i)×count(w j)Theδis used as a discounting coefﬁcient and prevents too many phrases consisting of very infre-quent words to be formed.The bigrams with score above the chosen threshold are then used as phrases.Typically,we run2-4passes over the training data with decreasing threshold value,allow-ing longer phrases that consists of several words to be formed.We evaluate the quality of the phrase representations using a new analogical reasoning task that involves phrases.Table2shows examples of theﬁve categories of analogies used in this task.This dataset is publicly available on the web2.4.1Phrase Skip-Gram ResultsStarting with the same news data as in the previous experiments,weﬁrst constructed the phrase based training corpus and then we trained several Skip-gram models using different hyper-parameters.As before,we used vector dimensionality300and context size5.This setting already achieves good performance on the phrase dataset,and allowed us to quickly compare the Negative Sampling and the Hierarchical Softmax,both with and without subsampling of the frequent tokens. The results are summarized in Table3.The results show that while Negative Sampling achieves a respectable accuracy even with k=5, using k=15achieves considerably better performance.Surprisingly,while we found the Hierar-chical Softmax to achieve lower performance when trained without subsampling,it became the best performing method when we downsampled the frequent words.This shows that the subsampling can result in faster training and can also improve accuracy,at least in some cases.Dimensionality10−5subsampling[%]30027NEG-152730047Table3:Accuracies of the Skip-gram models on the phrase analogy dataset.The models were trained on approximately one billion words from the news dataset.HS with10−5subsamplingLingsugurGreat Rift ValleyRebbeca NaomiRuegenchess grandmasterVietnam+capital Russian+riverkoruna airline Lufthansa Juliette Binoche Check crown carrier Lufthansa Vanessa Paradis Polish zoltyﬂag carrier Lufthansa Charlotte Gainsbourg CTK Lufthansa Cecile De Table5:Vector compositionality using element-wise addition.Four closest tokens to the sum of two vectors are shown,using the best Skip-gram model.To maximize the accuracy on the phrase analogy task,we increased the amount of the training data by using a dataset with about33billion words.We used the hierarchical softmax,dimensionality of1000,and the entire sentence for the context.This resulted in a model that reached an accuracy of72%.We achieved lower accuracy66%when we reduced the size of the training dataset to6B words,which suggests that the large amount of the training data is crucial.To gain further insight into how different the representations learned by different models are,we did inspect manually the nearest neighbours of infrequent phrases using various models.In Table4,we show a sample of such comparison.Consistently with the previous results,it seems that the best representations of phrases are learned by a model with the hierarchical softmax and subsampling. 5Additive CompositionalityWe demonstrated that the word and phrase representations learned by the Skip-gram model exhibit a linear structure that makes it possible to perform precise analogical reasoning using simple vector arithmetics.Interestingly,we found that the Skip-gram representations exhibit another kind of linear structure that makes it possible to meaningfully combine words by an element-wise addition of their vector representations.This phenomenon is illustrated in Table5.The additive property of the vectors can be explained by inspecting the training objective.The word vectors are in a linear relationship with the inputs to the softmax nonlinearity.As the word vectors are trained to predict the surrounding words in the sentence,the vectors can be seen as representing the distribution of the context in which a word appears.These values are related logarithmically to the probabilities computed by the output layer,so the sum of two word vectors is related to the product of the two context distributions.The product works here as the AND function:words that are assigned high probabilities by both word vectors will have high probability,and the other words will have low probability.Thus,if“V olga River”appears frequently in the same sentence together with the words“Russian”and“river”,the sum of these two word vectors will result in such a feature vector that is close to the vector of“V olga River”.6Comparison to Published Word RepresentationsMany authors who previously worked on the neural network based representations of words have published their resulting models for further use and comparison:amongst the most well known au-thors are Collobert and Weston[2],Turian et al.[17],and Mnih and Hinton[10].We downloaded their word vectors from the web3.Mikolov et al.[8]have already evaluated these word representa-tions on the word analogy task,where the Skip-gram models achieved the best performance with a huge margin.Model Redmond ninjutsu capitulate (training time)Collobert(50d)conyers reiki abdicate (2months)lubbock kohona accedekeene karate rearmJewell gunﬁreArzu emotionOvitz impunityMnih(100d)Podhurst-Mavericks (7days)Harlang-planning Agarwal-hesitatedVaclav Havel spray paintpresident Vaclav Havel graﬁttiVelvet Revolution taggers/p/word2vecReferences[1]Yoshua Bengio,R´e jean Ducharme,Pascal Vincent,and Christian Janvin.A neural probabilistic languagemodel.The Journal of Machine Learning Research,3:1137–1155,2003.[2]Ronan Collobert and Jason Weston.A uniﬁed architecture for natural language processing:deep neu-ral networks with multitask learning.In Proceedings of the25th international conference on Machine learning,pages160–167.ACM,2008.[3]Xavier Glorot,Antoine Bordes,and Yoshua Bengio.Domain adaptation for large-scale sentiment classi-ﬁcation:A deep learning approach.In ICML,513–520,2011.[4]Michael U Gutmann and Aapo Hyv¨a rinen.Noise-contrastive estimation of unnormalized statistical mod-els,with applications to natural image statistics.The Journal of Machine Learning Research,13:307–361, 2012.[5]Tomas Mikolov,Stefan Kombrink,Lukas Burget,Jan Cernocky,and Sanjeev Khudanpur.Extensions ofrecurrent neural network language model.In Acoustics,Speech and Signal Processing(ICASSP),2011 IEEE International Conference on,pages5528–5531.IEEE,2011.[6]Tomas Mikolov,Anoop Deoras,Daniel Povey,Lukas Burget and Jan Cernocky.Strategies for TrainingLarge Scale Neural Network Language Models.In Proc.Automatic Speech Recognition and Understand-ing,2011.[7]Tomas Mikolov.Statistical Language Models Based on Neural Networks.PhD thesis,PhD Thesis,BrnoUniversity of Technology,2012.[8]Tomas Mikolov,Kai Chen,Greg Corrado,and Jeffrey Dean.Efﬁcient estimation of word representationsin vector space.ICLR Workshop,2013.[9]Tomas Mikolov,Wen-tau Yih and Geoffrey Zweig.Linguistic Regularities in Continuous Space WordRepresentations.In Proceedings of NAACL HLT,2013.[10]Andriy Mnih and Geoffrey E Hinton.A scalable hierarchical distributed language model.Advances inneural information processing systems,21:1081–1088,2009.[11]Andriy Mnih and Yee Whye Teh.A fast and simple algorithm for training neural probabilistic languagemodels.arXiv preprint arXiv:1206.6426,2012.[12]Frederic Morin and Yoshua Bengio.Hierarchical probabilistic neural network language model.In Pro-ceedings of the international workshop on artiﬁcial intelligence and statistics,pages246–252,2005. [13]David E Rumelhart,Geoffrey E Hintont,and Ronald J Williams.Learning representations by back-propagating errors.Nature,323(6088):533–536,1986.[14]Holger Schwenk.Continuous space language puter Speech and Language,vol.21,2007.[15]Richard Socher,Cliff C.Lin,Andrew Y.Ng,and Christopher D.Manning.Parsing natural scenes andnatural language with recursive neural networks.In Proceedings of the26th International Conference on Machine Learning(ICML),volume2,2011.[16]Richard Socher,Brody Huval,Christopher D.Manning,and Andrew Y.Ng.Semantic CompositionalityThrough Recursive Matrix-Vector Spaces.In Proceedings of the2012Conference on Empirical Methods in Natural Language Processing(EMNLP),2012.[17]Joseph Turian,Lev Ratinov,and Yoshua Bengio.Word representations:a simple and general method forsemi-supervised learning.In Proceedings of the48th Annual Meeting of the Association for Computa-tional Linguistics,pages384–394.Association for Computational Linguistics,2010.[18]Peter D.Turney and Patrick Pantel.From frequency to meaning:Vector space models of semantics.InJournal of Artiﬁcial Intelligence Research,37:141-188,2010.[19]Peter D.Turney.Distributional semantics beyond words:Supervised learning of analogy and paraphrase.In Transactions of the Association for Computational Linguistics(TACL),353–366,2013.[20]Jason Weston,Samy Bengio,and Nicolas Usunier.Wsabie:Scaling up to large vocabulary image annota-tion.In Proceedings of the Twenty-Second international joint conference on Artiﬁcial Intelligence-Volume Volume Three,pages2764–2770.AAAI Press,2011.。

基于扩展Petri网的动态服务聚合流程描述模型及其BPEL4WS表示方法

ｕｃｒｉｃｉｔｐｉｎａｄｄ，ｎｃｖｒａｏｅｖｃｒｃｓｄｅｎｆｃｉｌ．Ｆｒｅｍｏｅｅｃｉｔｎｏｅｍｅｏｓｔｎｅｔｎａｔｉｏｔｎｙａｖｙｏ￣ｉａｉｔｎｏｓｒｉｅｉｐｏｅｓｍｏｌｅｅｔｅｙｕｔｒｒ，ａｄｓｒｉｔｔｄｏｉｆｎｉｇｖｈｐｏｆｈｈｔｎｌｔＷＳＰＲｎｔｎｏＢＥＡＷＳｗｓｍａｅｉａｌｔａｐｃｔｎＷｌｓｒｔｄｉｉｍｅｇｎｙｄｓｏａ．ｒｓａＣ／－ｅｔＰＩａｅｉａｄ．Ｆｎｌｙ，ｉｐｌａｉａｉｕｔｅｎｃｔｅｒｅｃｉｐｓ１ｓｉｏｓｌａｙＫｅｒｓｓｒｉｅｃｍｐｓｏ；ｒｃｓｏｅｎ；ＳＰＲｅ；ＢＥＡＷＳｙｗｏｄ：ｖｏｏｉｎｐｏｅｓｍｅｃｉｔｄｌｇＷＣ／ｎｔＰＩｉ
摘要：现有服务聚合流程建模方法的不足，于扩展Ｐｔ网提出了一种新的服务聚合流程／源针对基ｅｉｒ资
描述模型ＷＳＰＲｎｔ有效解决了动态服务聚合流程模型中不确定路径选择和服务的动态变化性问题。给出Ｃ／模型向ＢＥＡＰＩＷＳ的转换算法，以城市应急处理为例说明了转换算法的有效性。并
流程驱动（ｒｅｓｒｅ）ｐｏｓ。ｉｎ的服务聚合方法作为一种有效的聚合策略得到了很多研究者的关注。ｃｄｖ】ｊ流程建模是流程驱动的服务聚合的一个核心问题，目的是以一种形式化的方法对聚合流程进行直观其的描述和形式化的表达，从而为聚合流程的结构和性能分析提供技术基础。目前，内外很多研究人员基于基本Ｐｔ网和工作流网（一ｅ）的研究成果对服务聚合流程国ｅｉｒｎｔｊ建模技术进行了研究ｊ传统工作流应用中活动具有相对稳定性不同，ｂ务是动态变化的，。与Ｗｅ服因此，流程建模时不仅要保证任务流的畅通，聚合还要保障每一个流程环节上资源实现（ｂ服务）Ｗｅ的有效

微生物外文翻译之三

2. Growth on PAHs as sole carbon sources
Microbial degradation of PAHs and other hydrophobic substrates is believed to be limited by the amounts dissolved in the water phase (Ogram et al., 1985; Rijnaarts et al., 1990; Volkering et al., 1992; Volkering et al., 1993; Harms and Bosma, 1997; Bosma et al., 1997), with sorbed, crystalline, and non-aqueous phase liquid (NAPL)-dissolved PAHs being unavailable to PAH-degrading organisms. Bioavailability is considered a dynamic process, determined by the rate of substrate mass-transfer to microbial cells relative to their intrinsic catabolic activity (Bosma et al., 1997; Harms and Bosma, 1997). It has been described by a bioavailability number, Bn, (Koch, 1990; Bosma et al., 1997), which is a measure of a microorganism’s substrate degradation eﬃciency in a given environment. Bn is deﬁned as the capacity of an organism’s or a population’s environment to provide a chemical, divided by the capacity of the organism or population to transform that chemical. At high mass transfer rates, the overall biodegradation rate is controlled by the metabolic activity of the bacteria (Bn > 1), i.e. by both the speciﬁc activity of the cells and the population density. At Bn ¼ 1, the biodegradation rate is equally controlled by the physical transport and the microbial activity. When the transport of the substrate decreases or the bacterial population grows, the mass transfer becomes the factor that limits the biodegradation (Bn ! 1).

基于Petri网业务流程建模及到BPEL4WS的转换

ｐｏｉｒｌｘ，ＳｔｉｎｔｆｏｅａｓｒｃｄｌｇｏｏｉｅｓｐｏｅｓｈｒｆｒｈｒｓａｉｇｆｏｏｅｂｔａｔｍｏｅｓｔＥＬｉｅｙｏｉｓｏｔｆｒｔｂｔｔｍｏｅｉｆｂｓｓｒｃｓ．ＴｅｅｏｅｔｅｔａｌｔｒｍｔｒａｓｃｄｌｏＢＰｓｖｒｉｈａｎｎｎｎｈｒｓｇｉｃｔＢａｅｈｅｐｓｄｎｐｔｅｓａｄＢＰｉｎｆａ．ｉｎｓｄｔｅｄｅｔｙｏｅｒｎｔｎＥＬｇａｅ，ｔｉｐｅｅｎｓｓｖｒｌｏｏｅｔｆｔｅｐｔｉｎｔｄｔｅｔｉｓｕｉｌｕｇｎａｈｓｐａｒｄｆｅｅｅａｍｐｎｎｓｏｅｒｅｓａｒｌｉｃｈｎｈａｌｉＥＬｃｄｏｒｓｏｄｎｌｎｄｔｅｕｓｆｒｒｎａｇｒｔｍｂｕｏｔｒｓａｅｔａｎｇｔＢＰｏｅｃｒｅｐｎｉｇｙ，ａｈｎｐｔｏｗａｄａｌｏｉｏｈａｏｔｈｗｏｔａｌｔ．Ｆｉａｌｎｎｌｙ，ａｓｍｐｅｅａｌｓｎｒｉｌｘｍｐｅｗａｉｔｏｄｃｄｆｒｉｌｔａｎｏｔｅｔｅａｇｒｔｍｎｂｓｎｓｐｌａｏｉｈａｓａｅｐｏｅｅｅｆｃｉｅｅｓｏｅｍｅｏ．ｕｅｕｓｔｇｈｗｏｕｓｈｌｏｉｏｌｒｉｈｉｕｉｅｓａｐｉｔｎｗｈｃｌｏｃｂｒｖｄｔｆｅｔｖｎｓｆｔｔｄｃｉｎｈｈｈＫｅｒｓ：ＢＰｙｗｏｄＥＬ；ｅｒｎｔｂｓｅｓｐｏｅｓｍｏｌｇｐｔｅ；ｕｉｓｒｃｓ；ｄｅｉｉｎｎ

Atlas-based probabilistic fibroglandular tissue segmentation in breast MRI

1. Med Image Comput Comput Assist Interv. 2012;15(Pt 2):437-45.Atlas-based probabilistic fibroglandular tissue segmentation in breast MRI.Wu S, Weinstein S, Kontos D.Computational Breast Imaging Group, Department of Radiology, University of Pennsylvania, Philadelphia, PA, USA. shandong.wu@In this paper we propose an atlas-aided probabilistic model-based segmentation method for estimating the fibroglandular tissue in breast MRI, where a novel fibroglandular tissue atlas is learned to aid the segmentation. The atlas represents a pixel-wise likelihood of being fibroglandular tissue in the breast, which is derived by combining deformable image warping, using aligned breast contour points as landmarks, with a kernel density estimation technique. A mixture multivariate model is learned to characterize the breast tissue using MR image features, and the segmentation is subsequently based on examining the posterior probability where the learned atlas is incorporated as the prior probability. In our experiments, the algorithm-generated segmentation results of 10 cases are compared to the manual segmentations, verified by an experienced breast imaging radiologist, to assess the accuracy of the algorithm, where the Dice's Similarity Coefficient (DSC) shows a 0.85 agreement. The proposed automated segmentation method could be used to estimate the volumetric amount of fibroglandular tissue in the breast for breast cancer risk estimation.PMID: 23286078 [PubMed - indexed for MEDLINE]。

基于Petri网的建模技术ppt课件

• 从建模角度——可视化图形描述却被形式化数学方法支持；
8
Petri网建模的缺点： • Petri网的优点实际上是在模型构成上增加了模型的组成
元素，因此往往导致组成模型的元素数量过多； • Petri网不如基于活动网络容易理解； • Petri网的建模中不能在网中体现数据流，尽管基于状态
建模的Petri网能够精确、方便地对过程的控制逻辑进行定义，在这种情况下，数据流就与控制流完全混合，当两者不一样的时候， Petri网就无法显式地表示这种独立于控制流之外的控制流；
0
D=0 D=0
finish
ready
38
包含时间属性的交通灯
0
red1
0
30
safe
0
0
yr1
0
red2
30
yr2
rg1
yellow1
5
25
gy1
green1
yellow2
5
gy2
rg2
25
green2
39
层次的扩展
• 对复杂的Petri网添加结构信息的方法，与 DFD类似
• 一个子网是对库所，转
read_mail
send_mail
Hale Waihona Puke ready• 画出可达图 • 多少个可达状态? • 有无死状态? • 两个作者和三个读者的情况是怎样的?
32
agenda
➢ 1 Petri Net概述 ➢ 2. 经典Petri Net ➢ 3. 高阶Petri网 ➢ 4. 一个Petri网建模实例 ➢ 5.小结
12
Petri网的规则
• 连接是有方向的，其上可以标出权重 • 两个库所或转移之间不允许有边，且不应该有孤

BP神经网络模型

BP网络旳原则学习算法
BP算法直观解释
◦ 情况一直观体现
◦ 当误差对权值旳偏导数不小于零时，权值调整量为负，实际输出不小于期望输出，权值向降低方向调整，使得实际输出与期望输出旳差降低。
e
who
e w ho
>0，此时Δwho<0
BP网络旳原则学习算法
BP算法直观解释
◦ 情况二直观体现
xx1,x2, ,xn
h h y yo o i i h y h y o o ii1 1 1 1 ,,,,h h y y o o ii2 2 2 2 ,,,,
,h ip
,h o p ,yiq
,y o q
dod1,d2, ,dq
BP网络旳原则学习算法
◦ 输入层与中间层旳连接权值: w ih ◦ 隐含层与输出层旳连接权值: w h o ◦ 隐含层各神经元旳阈值: b h ◦ 输出层各神经元旳阈值: b o ◦ 样本数据个数: k1,2, m ◦ 激活函数: f ( )
将误差分摊给各层旳全部单元－－－各层单元旳误差信号
修正各单元权值
•学习旳过程：
• 信号旳正向传播向传播
误差旳反
BP网络旳原则学习算法-学习过程
•正向传播：
• 输入样本－－－输入层－－－各隐层－－－输出层
•判断是否转入反向传播阶段：
• 若输出层旳实际输出与期望旳输出（教师信号）不符
•误差反传
第七步，利用隐含层各神经元旳 h ( k ) 各神经元旳输入修正连接权。
和输入层
wih(k)weihhihe(k)hiw h(ihk)h(k)xi(k) wiN h1wiN hh(k)xi(k)
BP网络旳原则学习算法

人工神经网络建模matlab

• （5）神经网络可以用大规模集成电路来实现．如美国用 256个神经元组成的神经网络组成硬件用于识别手写体的邮政编码．
四、反向传播算法（B-P算法）
• Back propagation algorithm
• 算法的目的：根据实际的输入与输出数据，计算模型的参数（权系数）
• 1．简单网络的B-P算法
u (i) • (2) 设 k
表示第k层第i神经元所接收的信息
wk(i,j) 表示从第k-1层第j个元到第k层第i个元的权重，
ak (i) 表第k层第i个元的输出
• （3）设层与层间的神经元都有信息交换（否则，可设它们之间的权重为零）；但同一层的神经元之间无信息传输．
• (4) 设信息传输的方向是从输入层到输出层方向；因此称为前向网络．没有反向传播信息．
• 分类结果：(1.24，1.80)，(1.28，1.84)属于Af类；(1.40，2.04）属于 Apf类．
图2 分类直线图
• •缺陷：根据什么原则确定分类直线？
• 若取A=(1.46,2.10), B=(1.1,1.6)不变，则分类直线变为 y=1.39x+0.071
分类结果变为： (1.24,1.80)， (1.40,2.04) 属于Apf类； (1.28,1.84)属于Af类
• 问：如果抓到三只新的蚊子，它们的触角长和翼长分别为(l.24,1.80); (l.28，1.84)；（1.40，2.04）．问它们应分别属于哪一个种类？
• 解法一：
• 把翼长作纵坐标，触角长作横坐标；那么每个蚊子的翼长和触角决定了坐标平面的一个点.其中 6个蚊子属于 APf类；用黑点“·”表示； 9个蚊子属 Af类；用小圆圈“。”表示．

1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。

A Petri net semantic for BPEL4WS–validationand applicationKarsten Schmidt and Christian StahlHumboldt–Universit¨a t zu BerlinInstitut f¨u r InformatikD–10099BerlinAbstract.We translated a small business process into a recently deﬁnedPetri net semantic.Then we used the tool LoLA for validating the semanticas well as for proving relevant properties of the particular process.1IntroductionThe Business Process Execution Language for Web Services(BPEL for short) [CGK+03]is a language describing a stand alone business process as well as the composition of business processes.It makes a syntax available to model business processes based on Web services.This language is very expressive,but three lacks are known:First understanding BPEL is diﬃcult,because“it oﬀers(too)many overlapping constructs”[WADH02].Second BPEL builds on IBM’s WSFL[Ley01] and Microsoft’s XLANG[Tha01]and combines the features of both languages.That is the reason why BPEL has some st a mathematically sound se-mantic for BPEL does not exist.Thus,formal analysis of in BPEL speciﬁed business processes is impossible.BPEL is already used by business process developers.Accordingly it is necessary toﬁnd the language’s inconsistencies and make tool support available for verifying such processes.The main problem is the lack of a formal BPEL semantic.Most approaches ignore complicated but important concepts of BPEL like compensation and fault handlers.We have developed a pattern-based Petri net semantic for BPEL[Sta04].There-fore we can translate every business process speciﬁed in BPEL into a Petri net.Such a Petri net can be analyzed by existing tools.In order to validate the semantic and to verify translated BPEL processes we use the model checker LoLA[Sch00a].2Petri net semantic for BPEL4WSOur goal is to translate every BPEL process into a Petri net.Therefore the trans-lation is guided by the syntax of BPEL.Looking at BPEL,every process is built by plugging language constructs together.Therefore we translate each construct of the language into a Petri net.Such a net forms a pattern of the respective BPEL construct.Each pattern has an interface for joining it with other patterns as is done with BPEL constructs.Some of the patterns are used with a parameter,e.g.,there are some constructs that have inner constructs.The respective pattern must be able to carry any number of inner constructs as its equivalent in BPEL can do.We aim at keeping all properties of the constructs in the patterns.The collection of patterns forms our Petri net semantic for BPEL.In the following subsections,we give a glimpse on our semantic,using a basic activity(invoke)and a structured activity(ﬂow)as examples.2.1Example of a basic activityLet us have a more detailed look at the general pattern’s design.Figure1depicts the pattern for the BPEL’s asynchronous invoke construct1.Invoke is responsible for sending a request to a partner.In the asynchronous version,it does not wait for a result.In arc inscriptions of nets shown throughout this paper,a variable with small letter,e.g.,x symbolizes a single variable and a variable with a capital letter,e.g.,variables.X symbolizes a tupel ofIn general,a pattern is framed by a dashed box.Inside the frame,the structure of the corresponding BPEL construct is modelled.The interface is established by the nodes depicted directly on the frame.Controlﬂows from top to bottom while communication between processesﬂows horizontally.Outside the frame,there are external objects which relate to the scope,e.g.,obj1.The label on the top of an object deﬁnes its sort whereas the role is deﬁned at the bottom of the object.A sort is a set of tokens lying on and arriving at a place.The object’s role is independent of its sort.The pattern shown in Figure1reads a message saved in a variable and either invokes another BPEL process by sending this message or a fault is thrown because of a mangled message or some other error.Looking at theﬁgure the meaning of place stop,stopped and failed needs to be explained.In BPEL,a process is forced to stop its controlﬂow,e.g.,when a fault occurs or activity terminate is activated.However,the speciﬁcation[CGK+03]does not tell how to do so.Thus we have to make some modelling decisions in our model: 1In BPEL,an invoke activity may be conﬁgured such that it turns an idle process active.This is not shown in our picture.The pattern of BPEL’s scope is extended by a stop pattern,which has no equivalent construct in BPEL.If a scope needs to be stopped,the stop pattern controls this procedure.Our idea is to remove all tokens from the patterns,embedded in the scope pattern;thus the patterns of BPEL’s activities and compensation handler contain a subnet–a so called stop component.In the case of Figure1the stop component is established by transitions t4–t7using the interface stop and stopped.In order to explain how a stop component works imagine a scope that contains just an asynchronous invoke and the latter throws a fault.This leads to place failed being marked–the token is an object that consists of the fault’s name.This place is joined with a place in the stop pattern;thus this pattern gets the control of the scope.First it stops the inner activity of the scope and consequently a token is produced on the asynchronous invoke’s stop place.Transition t6ﬁres and stopped is marked.This place is also joined with a place in the stop pattern.If controlﬂows in a pattern and stop is marked,the transitions of the stop component are in conﬂict to them.This is in fact unavoidable,due to the involved concurrency.In[Sta04]we proved that using stop components every process can be stopped.2.2Example of a structured activityNext we show the general pattern of BPEL’sﬂow.Flow is used to execute subtasks concurrently.The subtasks can be further synchronized by so-called links.The pattern in Figure2can carry a number of n inner activities which are executed concurrently.Such an embedded activity can be any BPEL construct;thus only the interface is visualized and all other information of the pattern are faded out.Therefore only the frame and places initial,ﬁnal,stop,stopped and if neededFig.2.Pattern for BPEL’sﬂow embeds n inner activities.There are two possible sce-narios:Either all inner activities are executed concurrently(t2)and afterwards they are synchronized(t3)or the status of all source links embedded in theﬂow is set to negative (t1).negLink are visible(see innerActivity1in Figure2).The interface of each embedded pattern is joined with the surroundingﬂow pattern.negLink is an abbreviation of negative link.It is an optional place that is only part of a link pattern’s or of a structured activity pattern’s interface when it embeds at least one activity that is source of a link.With the help of negLink the status of all source links of an inner activity that are not executed anymore is set to negative, e.g.,imagine an activity within a branch that is not taken in a switch activity. In other words,negLink is a place for modelling Dead-Path-Elimination[LR99].In Figure2we assume that innerActivity1and innerActivityn contain at least one activity that is source of a link.If there is a token on stop,theﬂow and its embedded activities are stopped.After t5hasﬁred,the token lying on running is consumed;thus t3cannot be activated. Furthermore the stop place of each inner activity is marked.So innerActivity1,..., innerActivityn can be stopped concurrently.Firing t6synchronizes them.3ValidationFor validating the semantic,we generated a business process and veriﬁed several properties concerning the correct interplay between diﬀerent patterns.We use the Petri net based model checker LoLA that features powerful reduction techniques like symmetries[Sch00b]partial order reduction using stubborn sets[Sch99]and the sweep-line method[Sch04].In Figure3our example process is depicted–a modiﬁcation of the purchase order process given in the speciﬁcation[CGK+03,pp.14].A box frames an activity. In the case of a scope(see the box around the synchronous invoke)or the process itself we use a bold frame.Sequentialﬂow is depicted by dashed arcs,whereas concurrent activities are grouped in parallel.Arcs with solid lines symbolizes links.Fig.3.When the purchase order process receives an order from a customer,it initiates three tasks concurrently:calculating theﬁnal price for the order(left sequence),selecting a shipper(middle sequence),and scheduling the production and shipment for the order (right sequence).There are two dependencies between the three tasks,realized by links. After the completion of the tasks the invoice is sent to the customer.This is a small example,yet most of BPEL’s activities including fault handler and links are used.This process is modiﬁed by enclosing the synchronous invoke, which can send a fault message within a scope.This helps us validating especially the concept of stop in nested scopes.Furthermore we decided to abstract from data, i.e.messages are modelled as black tokens,because we directed our attention to the controlﬂow.The Petri net of the example process consists of158places and249transitions. It was generated manually.The whole state space consists of9991states and is calculated in less than a second.LoLA’s state space reduction techniques,in par-ticular partial order reduction and the sweep-line method,work well on the net. Applied together,state space is reduced to1286states.The results presented that the Petri net of a transformed BPEL process has too many places and transitions to represent it graphically,and given further that the nets behave well w.r.t.reduction techniques,deriving conjectures about the behavior of the net and model checking them turns out to be a reasonable technique for validating the semantic.Weﬁrst checked for dead transitions.LoLA found101such -paring these results to our patterns,we veriﬁed that exactly the right transitions are dead:Every pattern covers the whole behavior including all special cases of the respective BPEL construct.Consequently,in the case of such a small and simple process like in Figure3its patterns has some overhead not needed for the respective process.For a further use of the semantic,it would therefore be beneﬁcial to reduce the stuﬀnot necessary in a particular context away before doing veriﬁcation.Likewise,LoLA found15dead places which was expected as well.Next,we computed the196terminal states of the system and found them to be expected end states of the system.Weﬁnally veriﬁed several temporal properties such as“stop leads to stopped”,“the target activity of a link occurs always after the respective source activity”and so on.With the exception of some errors that were due to the manual generation of the net,LoLA conﬁrmed all results.For approving the source/target property of links,LoLA needed between386and4614states.4Veriﬁcation of a BPEL processThe objective in the previous section was to conﬁrm that the semantic itself is plausible.The results,however,encouraged us to go one step further and show that computer aided veriﬁcation of a BPEL process,using its Petri net semantic is indeed possible.We were able to verify relevant properties like termination,“the customer will always get an answer”,and“he will always get the correct result unless an error occurs”.The latter properties are complex temporal logic properties.To date, LoLA has only little support for such properties,so the state space amounted to between8728and8840states.For termination,1286states were computed.5ConclusionThis paper summarizes our work of verifying a business process speciﬁed in BPEL. We introduced in our pattern-based Petri net semantic of BPEL.An approach to validate this semantic was presented next.For an example pro-cess transformed into a Petri net,we veriﬁed several conjectures about the behavior of the patterns.Our example is also used to prove properties of the process,like termination.Furthermore it is shown that such Petri nets have relatively small state space;and state space reduction techniques behave well.Thus use of a model checking tool like LoLA is feasible.Further work of our research group is an ongoing validation of the semantic by transforming further BPEL processes.In order to handle larger processes,we plan to build a parser.It should be mentioned that the deﬁnition of a Petri net semantic,in connection with a concurrent approach based on ASM[Fah04],enabled us to discover a number of inconsistencies and ambiguities in the informal speciﬁcation.The results have been incorporated into recent working drafts of the document.References[CGK+03]Curbera,Goland,Klein,Leymann,Roller,Thatte,and Weerawarana.Busi-ness Process Execution Language for Web Services,Version1.1.Technicalreport,BEA Systems,Interantional Business Machines Corporation,MicrosoftCorporation,May2003.[Fah04]Dirk Fahland.Ein Ansatz zur formalen Semantik der Business Process Execu-tion Language for Web Services mit Abstract State Machines.Studienarbeit,Humboldt-Universit¨a t zu Berlin,July2004.[Ley01]Frank Leymann.WSFL–Web Services Flow Language.IBM Software Group, Whitepaper,May2001./webservices/pdf/WSFL.pdf.[LR99] F.Leymann and D.Roller.Production Workﬂow–Concepts and Techniques.Prentice Hall,1999.[Rei85]W.Reisig.Petri Nets.Springer-Verlag,Berlin,Heidelberg,New York,Tokyo, EATCS Monographs on Theoretical Computer Science Edition,1985.[Sch99]Karsten Schmidt.Stubborn set for standard properties.Proc.20th Int.Conf.Application and Theory of Petri nets,LNCS1639,pages46–65,1999.[Sch00a]Karsten Schmidt.Lola–a low level analyser.In Nielsen,M.and Simpson,D.,editors,International Conference on Application and Theory of Petri Nets,LNCS1825,page465ﬀ.Springer-Verlag,2000.[Sch00b]Karsten Schmidt.How to calculate symmetries of petri nets.Acta Informatica 36,,pages545–590,2000.[Sch04]Karsten Schmidt.Automated generation of a progress measure for the sweep-line method.In Proc.10th Conf.Tools and Algorithms for the Constructionand Analysis of Systems(TACAS)2004,volume2988of LNCS,pages192–204.Springer-Verlag,2004.[Sta04]Christian Stahl.Transformation von BPEL4WS in Petrinetze.Diplomarbeit, Humboldt-Universit¨a t zu Berlin,April2004.[Tha01]Satish Thatte.XLANG–Web Services for Business Process Design.Mi-crosoft Corporation,Initial Public Draft,May2001.http://www.gotdotnet.com/team/xml wsspecs/xlang-c.[WADH02]Petia Wohed,Wil M.P.van der Aalst,Marlon Dumas,and Arthur H.M.ter Hofstede.Pattern based Analysis of BPEL4WS,2002.。