Research And Realization Of Image Compression Algorithm Based On Wavelet Transform And ROI Coding

In the information age, people’s requirement of information is more and more big. Since the visual and understandable characteristics of image, it becomes the most important information source of human beings. However, one of disadvantages of image data is that the amount of it is very large, which will cause an enormous pressure to the storage and transmission. So the image compression technology has important research significance and application value.The goal of image compression is to achieve a high compression ratio under the premise that the quality of decoded image could not influence the demand. After a deep study of the existing compression technology, a novel compression algorithm which mixed lossy and lossless compression method is proposed in this paper. In the novel method, the wavelet transform is exploited to the image firstly, then lossless compression method is employed to the lowest frequency sub-band, which could retain most of information of the original image. The lossy compression method is employed to all of the high frequency sub-bands. Due to the sparse properties of the high frequency sub-bands, using the vector quantization with variable block techniques to carry out a high ratio of compression is possible. In addition, an initial codebook design method which is based on the energy sorting method is put forwarded. It improved the performance of the classical codebook training algorithm of LBG Otherwise, to reduce the coding time, a new codeword search algorithm which is based on elimination of abnormal codeword is proposed, the new code search method almost does’t introduce search distortion. Besides, the correlation of the same direction high frequency sub-bands is used to present a fast block-partitioning method. Furthermore, to meet the needs of some of special demand of some image, the lossless stream of the difference image of region of interest is appended to the ROI, and the lossless compression of region of interest is realized.The contract experiment between the proposed method and the traditional algorithm shows that the execution speed of the proposed method is faster than that of traditional algorithm. What’s more, the decoded image quality is superior to the traditional algorithm at high compression ratio. The proposed method is more suitable for high rate compression.