Intelligent Image Compression Scheme Based On Multi-scale Transform

With the development of Multimedia, the images become important as a media for transforming information. At the same time, more and more people attach great importance to image processing technology. In this paper, to accomplish an intelligent image compression scheme based on multi-scale transform, we studied the wavelet theory and transform coding. After three innovations, a novel static image scheme in accordance with mentioned demandes was completed. The first innovition is combines SVM regression with wavelet transform coding. We design an intelligent image compression scheme named W-SVM algorithm, which can improve the compression ratio effectively. Although the compression performance of W-SVM algorithm gains much improvement, the reconstructed image has shown block effect and blurred edges. To improve W-SVM algorithm, we introduced curvelet transform into the new image compression scheme. Curvelet transform is one of the multi-scale transforms, which represent singular curve sparsely because of oblongshaped basis function. And the second innovition named C-SVM algorithm works fairly well for declining block effect at higher compression ratios. On the other hands, C-SVM algorithm has disadvantages. For example, C-SVM algorithm spends a lot of time in performing. To reduce the total performing time, we design an image compression schem based on curvelet transform and CVM (core vector machine), which can be called C-CVM algorithm. Compared with image compression algorithms do not use CVM and methods based on wavelet transform, experimental results show that the compression performance of our method gains much improvement in peak-signal-to-noise-ratio (PSNR) and CPU time. Moreover, the algorithm works fairly well for declining block effect at higher compression ratios.