基于小波系数子树的分形图像数据压缩

目的 研究小波变换和分形编码在图像数据压缩中的应用。方法 根据小波变换后不同分辨率下图像的子带小波系数的相似性，通过分形非收缩仿射变换，利用低一级频率分辨的子带小波系数子树预测第一级频率分辨的子带小波系数子树，实现对图象的高效描述。

Fractal Image Data Compression Based on Wavelet Coefficient Subtrees

Aim To study the application of wavelet transformation and fractal coding in image compression.Methods\ Due to the self similarity existing among the wavelet coefficients of the subbands with different resolutions and spatial orientations, fractal noncontractive affine mapping transformation was used to predict higher resolution subband wavelet coefficient subtrees with lower resolution ones to fully exploit this self similarity, obtaining high efficient presentation of an image. Results\ Experiment results show that this scheme can obtain high compression ratio while keeping certain reconstruction image quality (e.g. P SNR = 29 08 dB and compression ratio 74 8). Conclusion\ This scheme can take full advantage of wavelet transformation coding and fractal coding and can provide a new insight on low bit rate image compression.