Fingerprint Enhancement

(a) (b)
Figure 2: (a) Original image (b) Enhanced fingerprint. 2.2 Implementation of Spatial Filtering The purpose of fingerprint enhancement filtering is to smooth (interpolate) the ridges along the same ridge line, characterize edges, and sharpened details in perpendicular of the
…….
Wavelet Transform Local Ridge Orientation Estimation
……. Texture Feature Extraction Enhancement Algorithm
Reference Point Determination
Minutiae Extraction
Sanpachai Huvanandana, Ph.D. and Settapong Malisuwan, Ph.D.
Department of Electrical and Computer Engineering Chulachomklao Royal Military Academy Nakhon-Nayok, Thailand
A Hybrid System for Automatic Fingerprint Identification:Fingerprint Enhancement
A Hybrid System for Automatic Fingerprint Identification: Fingerprint Enhancement
⎛ n+ P / 2 r + P / 2 ⎜ ∑ ∑ 2 J xij J yij −1 ⎜ i = n − P / 2 j = r − P / 2 = 0 .5 tan ⎜ n + P / 2 r + P / 2 2 2 −Jy ⎜ ∑ J x ij ⎜ ∑ ij ⎝ i=n− P / 2 j=r − P / 2 ⎞ ⎟ ⎟ ⎟. ⎟ ⎟ ⎠
International Journal of The Computer, the Internet and Management Vol. 13 No.2 (May-August, 2005) pp 1-12
1
Sanpachai Huvanandana and Settapong Malisuwan
Θ nr
(
)
(1)
Texture Code Construction
Minutiae Database
Fast Searching
Texture Matching Result
Minutiae Verification
No
Matching Score > th No Yes Identified person
Texture and fingerprint class database, and distance matrix.
Reject
Figure 1: The proposed hybrid fingerprint identification system. 2. Enhancement and Minutiae Extraction 2.1 Local Ridge Orientation (LRO) Estimation The optimal ridge direction is defined as the directional field that shows the dominant local structure orientation. We propose an orientation angle approximation via Wavelet Transform. Level-1 of a 2-D wavelet represents 4 sub-bands, called LL, HL, LH,
Fingerprint
and HH, which correspond to low frequency, horizontal detail, vertical detail, and diagonal detail coefficients, respectively. Let Θ be an orientation angle, Jy and Jx are the vertical (LH) and horizontal (HL) wavelet coefficients, respectively. The estimate of the dominant orientation Θ at the center (n,r) of P x P non-overlap block is given by,
Abstract A hybrid fingerprint identification system is presented in this paper. The system consists of several steps: fingerprint enhancement, minutiae extraction, texture feature extraction, fast texture matching, and minutiae matching. The enhancement algorithm consists of several steps: local ridge orientation estimation, filter design, enhancement, intensity stretching, binarization, thinning, minutiae extraction, and false minutiae elimination. Our approach has been tested in term of both objective and subjective evaluation. The experimental results show an accurate minutiae extraction and effective performance in fingerprint matching system. The proposed hybrid matching system needs fewer steps to identify individuality than traditional methods. The proposed hybrid matching system has been tested on a reasonably large fingerprint database and the experimental results show an effective performance without matching every fingerprint in database. 1. Introduction Among various fingerprint identification systems, the traditional methods based on minutiae have been widely used [1, 2.3] because the topological structure of minutiae of a fingerprint is unique and invariant with
The advantage of using wavelet coefficients for orientation angle estimation is that the LH and HL wavelet coefficients, which can be viewed as smoothed gradient information, can be directly calculated and treated as partial derivative results. Hence, smoothing is not required after orientation estimation. In our experiment, the proposed LRO estimation outperforms other gradient filters such as Sobel, Robert, DoG, and LoG.
aging. Jain et. al. proposed an novel alignment-based matching algorithm in [2] where the authors used ridges associated with minutiae to obtain transformation parameters and elastic matching had been used for matching. The result was accurate but the system required a large storage because all ridges information must be saved. Another fingerprint identification approach, called a filterbank-based fingerprint matching [4], used a bank of Gabor filters to capture the details in fingerprint as a compact fixed length FingerCode. The matching is based on Euclidean distance between two FingerCode, hence, it is faster than traditional method but reported a lower performance. In the filterbank-based approach, convolution with Gabor filters is the major computation to the overall feature extraction time. In this paper, we focus on a framework of an automatic fingerprint identification using a hybrid system with a fast filters-based searching and a minutiae matching. The flow chart of our proposed hybrid fingerprint identification system is shown in Figure 1. In Section 2, the fingerprint enhancement and minutiae extraction framework are explained. Section 3 discusses a proposed fast fingerprint matching system using a filterbank. A hybrid fingerprint identification system is proposed in Section 4.