The Research For The Image Contour Compression Method

The image includes the very abundant information, which looks like intuitionisticand lively, so that it becomes the important intermedium of transferring theinformation. At the same time, the huge data quantity is the big difficulty forcollecting, storing, processing and transferring, which is the important reason that theimage can't be the primary intermedium. Thereby, compressing the ratio of digitalimage signal becomes the chief task in the field of the image communications and theimage signal processing, which attracts the technological operator all over the world.The wavelet transform is the more prevalent and best compressing technology atpresent, which mostly acquires the good effect to compress the image texture andsmooth region, but it has the shortage to compress the contour ---the importantcharacter of the image. The research is finite to it at present, which mostly have theCART model of Donoho introducing and the improved model of Romberg et.al.Firstly, the article introduces the criterion of the compress technique and thetypical compress algorithms, also fully introduces and analyzes the developmentstatus of these algorithms in the world, and import the compress algorithm based onthe wedgelet. Secondly, it introduces the theory of the image Binarization and theprimary Binarization Algorithm presently. The Binarization aims to reduce theinfluence of the multilayer gray value for the image processing, and the correspondingexamination has been done to validate it. Thirdly, it introduces the theory of the imagesegmentation and frequent segmentation technique. The aim of the imagesegmentation is to mark off the interested region of the image, which is done to thecorresponding image processing to reduce the operation time and save the storagespace. At last, it mostly introduces the compress technique based on the wedgelet, andcompares the CART model which is introduced by Donoho and the improved modelby Romberg et.al. with the model of the article, then compares the image ofprocessing based on the three methods. It can be seen from the results that thealgorithm of the article effectively wipes off the disturb signal by the imageBinarization, introduces the representation model to reduce the calculating complexityand also reduce the calculating quantity, obtains the clear contour and acquires theupstanding effect.