The Design Of Algorithm Of Multi-Computers' Parallel Wavelet Three-dimensional Images Compression And It's Realization In LAN

In the current society, people's requisition for communication has already moved from simple symbols, characters, pronunciation communication towards symbols, characters, pronunciation, static image, video images, etc., namely the multimedia communication. Image communication includes mainly the processing of digital picture (including: sampling, scan, compression or decompression, store etc.) and transmission. In the past ten years, the coding technology of images has got fast development and extensive application, and it gets mature day by day. Its sign is the formulation of several international standards of image coding, including static coding standard JPEG, video telephone /video meeting standard H.261, video image standards MPEG-1 and MPEG-2, etc. These standards have merged various kinds of good performance traditional image coding methods at present. High-dimensional (spatial or video) image data are used increasingly in many fields, yielding huge datasets. To store and transmit these images, we need an effective compression method for high-dimensional images. 2D wavelet compression methods have been used as the basis of a new international standard for still 2D image compression (JPEG 2000). To improve the compression of high-dimensional image, we research on high-dimensional wavelet compression methods,in order to increase the compression ratio by exploiting the information coherence in higher dimensions. In this paper, we designed a 3D parallel wavelet compression algorithm. And we distributed the process of image encoding and decoding, which was done on a single computer in the traditional model, onto a multi-computer system based on local network, in order to reduce compression time. In this paper, we divided a three-dimensional matrix into two small blocks. One was sent to client computer to process and the other one was processed by host computer. The whole process included analysis, quantization, scanning, coding and their reverse process. At last, the two small blocks were combined into an integral unit by host computer. The process includes data transmitting in LAN, the application of MPI (Message Passing Interface), the distributed parallel computing etc. Experiments show: the given 3D wavelet compression method allows much higher data compression ratio than a comparable 2D wavelet compression method, and using less time in a multi-computer network than in a single computer, while remaining image quality to meet people's demands. The methods in this paper can be used for video image transmission in Internet and telemedicine, etc.