Arithmetic Research And Realization Of Static Image Compression Coding Base On JPEG2000 Standard

With the development of the multi-media and network technology, people has put forward the higher requirement at quality dimension and application in digital image, expecting to use limited space and bandwidth resources to transmit a big image and according to the real needs to gain the reconfiguration image of different resolution or qualities. This will need that the image compression technology not only having good compression efficiency but also having the capability of handling nimbly the compression code rate. At present, the JPEG stationary picture compressing standard has a good compressing performance at middle and high compressing rate. But there exists a serious diamond effect at low compressing rate. However, JPEG standard can not very well make the needs of transmitting mages by networks at low compressing rate. In order to solve the issue mentioned above, the international standard organization (ISO) has laid down the new generation stationary picture compression standard: The JPEG2000 (serial number is ISO-15444). This thesis analyses and studies ISO 15444-1 that has been formed the international standard.In this thesis, The theory and advantages of JPEG2000 and wavelet transform theory ara briefly presented firstly. Secondly, based on analyzing the characteristic of image compression using wavelet transform, both critical problems and generic steps of the wavelet transform used in image compression are discussed. Finally, the algorithm embeded zero-tree wavelet is analyzed and discussed in detail, and a improved algorithm based on zero-tree is proposed, which adjusts the structure of the zero-tree rationally based on the classic EWZ coding technology. It also changes the original coding flow, which results in the adjustment of processes of main scanning, assistant scanning, and decoding. And wavelet transform and EZW arithmetic are achieveed using C++ language.Experimental results show that the compression ratio of thezero-tree-based algorithm is far more advantage than that of the EZW-based one, and its codes can be intercepted at random without any additional distortion, and remains the high quality of the original image. Yet the novel zero-tree algorithm has some shortage to be improved.