A new image coding algorithm is presented, which combined the fractal and discrete wavelet transform, based on the statistical character of image blocks, the distance between the value block and its best matched domain block in the baseline fractal coding and the distribution of noise unknown. In the proposed algorithm, ifthe robust regression between the value block and it' s best matched domain block is less than the given threshold value, the value block is compressed by fractal coding, otherwise is compressed by discrete waveform transform. Simulation results show that the proposed algorithm can speed up encoding time greatly and improve on the quality of the reconstructed image. Especially that has good robustness against the outliers caused by salt and pepper noise.%针对图像易受外界噪声干扰,且这种噪声分布通常是未知的这一问题,结合图像的统计特性,基本分形编码中值域块和最佳匹配的定义域块之间的距离统计特性等,提出一种基于方差不变特性、邻域搜索的分形与小波相结合的图像分形编码算法.在该算法中,如果值域决和最佳匹配之间的稳健回归优化目标函数取值小于给定的阈值,则用分形压缩算法编码该块,否则用小波变换压缩该块.实验结果表明,该方法可使编码速度比基本分形算法有较大提高,而且原始图像在受到外界干扰的情况下,该算法表现出了较好的鲁棒特性.
展开▼
