Some Research Of Compressing Coding Algorithm For Binary Images

As the development of information technology, more effective storage and transmission of multimedia files, such as videos and images, are demanded, and this results in a higher request for data compression. Image coding is an expert technology that works on the compression of images data, and so far there are almost a hundred kinds of compressing methods, and also many international standards like JPEG, MPEG and JBIG In recent years, image coding technique has had great progress, and a lot of new compressing thought and ways were proposed, like fractal coding and wavelet transform coding. Oppositely, as an important embranchment of image compressing, the binary image coding has a more slow evolution, performed as low compression rate and singleness in coding way, which hasn't been enough for actually need. Binary images, which are also called bi-level images, are a specific kind of gray images. Although have simple configuration and small data quantity, they are very important both in usual life and many scientific research like image processing and pattern recognition. And text files, project graphics, fax, newspapers can be all seen as binary image, and besides, large quantity of binary images storage and processing is needed during the study of pattern recognition, object detection, movement tracking and medical images processing. Additionally, the development of binary image coding can be sure a good acceleration for gray ones. And in all, it will make great sense to keep studying coding of binary images. Now the primary methods are lossless coding, include White Block Skipping, Run-length Coding, Quadtree Coding and so on. The JBIG standard was confirmed by ISO in 1993. But, all these methods and standard have their own disadvantages more or less.In order to get higher compression rate, in this thesis, we have lots of study on WBS, RPC and Fractal Coding, and some improvements are proposed on the base of WBS, and a new lossy compression coding based on the area character of binary images is created also. The WBS makes compression only to the white pixels, and it can get great CR with images those are occupied almost wholly by white. But it dose not have good performance to most actual images with more black because of its excessively simple coding way. Two improvements are brought out against the limit of its singleness, which make the algorithm more general so that it can be used to more usual images.Analysis-integration is a kind of coding thought of images, in which all the image elements will be classified with some rule first, and then different methods can be used according to different category at the encoding side, and all parts will be integrated at the decoding side. In this thesis, after studying the pixel character of binary images, a new analysis-integration lossy coding method is proposed, in which the whole image is divided into two different areas, the center area and the edge area, and to the former, we use rectangular coding, to the latter, we use near-fractal coding. And, fractal coding is a compression method of gray images developed fast in last twenty years, which has very high CR. The attempt that combines fractal to binary image coding makes good result, proved by our experiments.