基于能量树编码的小波图像压缩算法

针对小波变换的空间能量聚集特性,提出了一种基于能量树编码的小波图像压缩算法。该算法在离散小波变换的基础上,分别对图像的各高频子带按其局部能量构建分层能量树,利用总能量和各层的能量角等效表示子带的小波系数;根据给定的压缩比,选择合适的代价函数构建最佳能量树,然后对其进行量化和编码,通过自适应的比特率分配实现小波图像压缩。实验结果表明,该算法实现简单,重构图像质量好,与当前多种主流的小波图像压缩算法相比,压缩性能有了明显提高。

Wavelet image compression based on energy-tree coding

According to the spatial energy aggregation properties of wavelet transform,a new wavelet image compression algorithm based on energy-tree coding is proposed.On the basis of discrete wavelet transform,the algorithm respectively constructs hierarchical energy-trees of all the high frequency sub-bands by direct measures of local energy,and equivalently represents sub-band coefficients with the parameters of its total energy and energy angles of all the levels.In light of the given compression ratio,select an appropriate cost function to construct the optimal energy-trees,then digitize and code their parameters,and the wavelet image compression is completed with adaptive bitrate allocating.Experimental results show that the proposed algorithm is of better compressing performance with simple implementation compared to many state-of-art wavelet coders.