Research On ID Photo Coding System Based On Shape Adaptive

For the traditional method of image compression, the entire image is compressed as anobject with the same manner. However, we tend to concentrate on some regions of an imageinstead of the whole image and we obtain the information from an image selectively. Thus weapply different compression quality in regions of interest and non-regions of interest, whichsatisfies the human vision system and improves the efficiency of image compression. Inrecent years, the image coding technology for arbitrary shape images attracts the attention ofmore and more people. This coding approach is to divide an image into different regionsbased on certain rules and apply corresponding coding schemes for different regions. IDphoto is a kind of images whose foreground and background are obviously distinctive. In thispaper, we introduce a coding algorithm of arbitrary shape images based on contour for thecompression coding of ID photos. The experimental results indicates that this method canimprove the compression efficiency, ensure the good visual effect of ID photos and have thecharacter that the foreground regions of arbitrary shape for ID photos can be accessedseparately.This paper includes the following five parts.1. The shape adaptive based coding system of ID photo is designed. This coding systeminvolves the finely segmentation of simple background image, shape adaptive discretewavelet transform, contour coding, zero-tree coding of arbitrary shape and adaptive arithmeticcoding etc.2. The method of finely segmentation for ID photos is designed. The ID photos aresegmented by combing quadtree splitting and merging with edge detection, which solves theproblem of hair area burr of ID photo, and improves the compression ratio of ID photo.3. Research on shape adaptive coding method based on contour. This kind of method isdifferent from the General-Scaling-Based method and the Max-Shift method in JPEG2000standard upon coding arbitrary shape region. The Advantage of this method is that the codingregions can be randomly stored, accessed and transported. The lack is that the prospectregion’s contour is needed coding.4. The shape adaptive based coding system of ID photo is simulated via Matlab2010on the Windows7operating system, including segmentation subsystem, contour codingsubsystem and prospect area coding subsystem.5. The contrast experiment has been done and the results have been analyzed. The resultsof proposed method is compared with JPEG, EZW, and SPIHT, and we conclude that thecompression effect is superior than other three methods when the PNSR(Peak Noise to SignalRatio) is less than30dB.