一种基于小波变换的图像压缩方法

提出一种基于小波变换的灰度图像数据压缩编码方法，基本思路是利用小波变换实现图像的多分辨分解，用矢量量化（ＶＱ）对分解后的图像进行编码，在矢量量化ＬＢＧ算法的初始码书的选取中根据矢量中各分量的特性提出一种改进的随机选取法，避免了可能的胞腔不均现象，提高了码书的质量，而且重构的图像质量也有所提高。

New Image Compression Scheme Based on Wavelet Transform

Based on wavelet transform and vector quantization, a scheme for image data compression is presented. Firstly, in this method, the image is decomposed in multiresolution by wavelet transform, and then the image is encoded by vector quantization(VQ). A new way,which is used to get the initial codebook of LBG algorithm in VQ, is put forward. The new method is based on some characteristics of components in the vector so as to improve the performance of the orignal random algorithm, and the codebook is more reasonable. A number of computer simulation experimental results show that the new method exhibits a better performance compared with Huffman and JPEG coder, and a high quality reconstructed image can be achieved in a low bit rate.