基于小波包矢量量化图像压缩方法

A New Image Copression Scheme Based on Wavelet Packet Vector Quantization

提出了一种新的数字图像压缩方案———基于小波包的区域联合矢量量化 (WPVQ) .WPVQ压缩方案充分利用了小波分析的三个重要特点 ,即零树特性、多层分解和适合跨区域联合编码 .在具体实现过程中 ,WPVQ又采用了一系列措施来提高信噪比并降低算法复杂度 .计算机模拟实验表明 ,同采用嵌入式零树编码 (EZW )的JPEG2 0 0 0相比 ,WPVQ可以在高压缩比情况下保持良好的信噪比和主观效果 。

A new image compression scheme WPVQ is proposed, which is based on wavelet packet and regional combining vector quantization. WPVQ scheme utilized three most important characteristics of wavelet transform include zerotree characteristic, multilevel wavelet decomposition and regional combining encoding. In the process of scheme realization, a series of measures are taken to enhance SNR and to reduce algorithm complexity. The simulation result shows that, cam pared with JPEG2000 which has adopted the Embedded Zerotree Wavelet algorithm (EZW), WPVQ can keep high PSNR and good subjective evaluation when CR is rather high, and especially, it's very suitable for high CR compression to digital images.