几种图像变换算法性能比较

Comparion of Servral Image Transform

为寻找提高图像压缩算法性能的途径,比较了二维DCT(Discrete Cosine Transform)变换,二维张量积小波变换以及最新的几何小波的变换特点及实用效果。针对同一图像采用不同的几何小波进行分解,保留相同个数的显著系数进行图像重建,以比较各种小波基的稀疏表示能力。结果显示,在高压缩比条件下,Ban-delet和DCT变换更加有效,而在高质量条件下,小波变换依旧是最有效的工具。

In order to find a way to improve the performance of image compression,the characteristic and effect of 2D-DCT （ Discrete Cosine Transform） ,two-dimensional tensor product wavelet,and the latest geometrical wavelet are compared. The same image is decomposed with different geometrical wavelet and then reconstructed with the same number of significant coefficients,thus the sparse representation capability of wavelet can be compared,the result shows that Bandelet and DCT are more efficient when require high compression ratio,while wavelet is still the most efficient tool when require high image quality.