LCD Motion Blur Blind Modeling and Analysis

(a) (b) Figure 1. Pixel Intensity vs. frame time for (a) LCD and (b) CRT displays. CRTs use short persistence phosphor, resulting in very short integration time and no motion blur [Fig. 1]. Large LCD panels produce more motion blur. Among other reasons, eyetracking is done more readily on larger screens, resulting in
I I
t
t
more spatio-temporal integration, therefore more motion blur, which was analyzed in [1]. In recent years, a lot of efforts have been put into speed the response time of LCD . Response Time compensation (RTC) has been successfully applied to reduce LCD response times [2]. With RTC, overdriving techniques reduces motion blur by enlarging the desired change in the pixel value in order to force the LCD material to react faster. Therefore RTC can reduce motion blur, resulting in enhanced motion image quality. However, Theoretically, even with a zero response time, the LCD will still emit approximately constant light through the frame period, hence causing motion blur. Recent studies on LCD motion blur calculated that the sample-andhold property of the LCD display contributes to 70% of the visible motion blur[3]. In order to reduce LCD hold-type time ,previous methods for reducing LCD motion blur include variations of flashing the backlight [4], frame rate up conversion[5,6], motion compensated inverse filtering (MCIF)[7],and Richardson-Lucy (RL) deconvolution[8] Although LCD motion blur has been addressed extensively in the literature, there are a few LCD motion blur models.In the past,several models have been reported in literature including “sinc”model[1] and “sin” model[3]. The most useful information when attempting lcd moton deblur is an accurate psf.(or model),however this can be difficult to obtain. In this paper, an effective blind modeling method for LCD motion blur were proposed, which obtaining LCD motion blur images by MPRT(Motion Picture Response Time) simulation method and modeling by blind deconvolution method. We are able to obtain a better model for LCD motion blur than “sinc “model and recover the sharpness of moving images on LCDs to a large extent. The paper is organized as follows. First, we will introduction LCD motion blur images obtaining by MPRT simulation method. Second, we will build a LCD motion blur model using iterative blind deconvolution algorithm. Third,we will verify this model in standard Richardson-Lucy (RL) deconvolution, followed by conclusions. II. SIMULATION LCD MOTION BLUR
A perceived blured motion image on LCD is not the same as a moving-picture photograph because of the human-vision characteristics and hold-type display affect. LCD Motion blur can be characterized by the motion-picture response-time metric (MPRT). MPRT values take into account the LCD
H.Z. Tan
School of Information Science and Technology, Sun Yat Sen University Guangzhou, China 510275 issthz@
Abstract—LCDs(Liquid Crystal Displays) have displaced cathode ray tube(CRT) displays in most applications, however motion blur is one of the main challenges for using LCDs in television applications. LCD motion blur is caused jointly by the slow LCD’s temporal response and the hold-type LCD’s rendering method together with the motion pursuing function and the temporal low-pass filter characteristic of human visual system(HVS). Although LCD motion blur has been addressed extensively in the literature, there are a few of LCD motion blur models. In this paper, we show an effective blind modeling method for LCD motion blur. Simulation and perceptual tests indicate that our model is indeed effective at reducing the amount of motion blur on LCDs . Keywords- LCD; motion blur; Motion Picture Response Time; blind model; blind deconvolution; point spread function
I.
INTRODUCTION
Recently, LCD displays are becoming more and more popular in the TV Market. Compared with the traditional CRT display, LCD is thin, flat, energy efficient, light-weighted etc. But CRT is still unbeaten in one major aspect–fast motion portrayal. LCD weaknesses like motion blur affects the TV watching experience drastically. LCD’s rendering method, called “hold-type” in the display industry, is different from CRT’s rendering method, called “impulse-type.” As illustrated in Figure 1, the hold-type LCD’s pixels are always on during a frame cycle, while the impulsetype CRT’s pixels are only on for a very short period during a frame cycle.