Quantum codes image compression algorithm yards book research and its decoding circuit the realization of

The vector quantization technology because of its algorithm is simple, easy to hardware realize, in the image compression dedicated chip-based designs is widely used. Vector quantization technology, the key is the feature image database the establishment and search algorithm select it. For an existing vector quantization technology in the hardware realization when the costs are high, cost-effective low, and the latency serious problems, the present thesis adopts improved algorithm's vector quantization technical (quantum codes), effectively raising the image compression decoding dedicated chip's compression rate and the encoding speed. Decoding circuit structure is simple, easy to implement. Experiments show that after decoding image picture quality is better, encoding speed has also a substantial increase in, allows ordinary communication network is real time image transmission become possible. Thesis take advantage of wavelet transform in handling two-dimensional discrete signal on the advantages of, combined with image's visual feature, by using a two wavelet transform analysis of image block vector of directionality of, will be code vector in accordance with the vertical direction of the, the horizontal direction, irregular Changes and gently varying is divided into four categories, in the each category's code vector middle extract the signature vector database (signatures book). Thereby the traditional vector quantization to use the feature database in accordance with the directivity divided into four, encoding when the first determines that the image block direction of the vector, and then in the corresponding database (classification Codes book) in the match. Thereby reducing the matching search time, while restoring image of the peak signal to noise ratio basic remain unchanged. Experiments show that, in PSNR decreased only 1.8% the case where, coding speed an average of by 38. 4%. The maximum can improve the 45.8%. Secondly, the proposed based on classification coding algorithm vector quantization coding system hardware structure, while simplifying the classification algorithm's judgment conditions for, thus reducing the circuit scale. And through the RTL level simulation, integrated wiring after the's doors-level simulation, indicating that the of the design the rationality Xing and stop indeed Xing. Once again, put forward PDVQ (partition dynamically VQ) algorithm of the decoding circuit of the structure, with Verilog-HDL completed a RTL-level code describes, and its function carry on verification, test results with the software simulation results completely anastomosis, indicating that the correctness of the design . Xi'an University of Technology master's degree thesis Finally, as a complete encoding decoding system, This thesis is use the RTL-level code describes the adopt the above-mentioned algorithm is's VQ compression decompression chip, the and verify the its functionality the integrity of the. Keyword vector quantization peak value signal to noise ratio wavelet transform classification yards books topics by the the National Natural Science Fund project funding, the fund No. 60,276,017