Compressing Image By Using Wavelet Analysis

With the popularization of computers and network technology, people brought forward higher requirements on the quality, size and application of digital images, hoping to store and transfer big images by using limited space and bandwidth, and to obtain images of various resolution ration or quality. This requires the image-compression technology to be not only more efficient in compressing images, but also to be more flexible in dealing with compress coding ratio. As a result of all the above, great attention has been paid to the discussion and research of code-compressing method in recent decades. During the latest 20 years, great progress has been made in the theory, method and technology of image-compression coding on the basis of such traditional coding methods as coding forecast and exchanger coding etc; wavelet transformation has been widely used in signal-disposal field, because it has a characteristic of energy focus, it puts forwards a concept of multi-resolution ration analysis and the realization of algorithm about wavelet decompose and fast rebuild. Among the applications of wavelet analysis, it is rather mature to apply it to image-data compressing. In the first place, this article introduces several traditional code-compressing methods, and makes a brief introduction about the international standard JPEG and JPEG2000, and dwells on the wavelet multi-resolutions dissecting, decompression and rebuild, EZW algorithm, improved SPIHT. Then, the idea to pre-treat images is put forward on the basis of what has just been discussed, combining with the visual characteristic that people are more sensitive to low-frequency than high-frequency parts of images, different coding algorithms are suggested to use for various frequency. During the image-compressing, lossless-compression method (Huffman-Coding) is used to deal with the low-frequency part, while SPIHT-coding is used to deal with the high-frequency part, so mat high compression scale can be achieved without worsening the SNR. In the article, the quality of the rebuilt images obtained by using different coding methods is compared. At last, the core technology of compressing moving images is discussed-motive reward technique, and the compression of moving images is realized by using the method that the two wavelet transformations of two-dimensional still images and one-dimensional time-axis are combined together.