Research on Image Compression Coding Based on Number Entropy´s Optimal Wavelet Packet Base Tree

The traditional method of image coding compression for wavelet transformation emphasized the low frequency signal, while ignored the high frequency, on which energy of images chiefly concentrates. It could lead to distortion of the image. Considering the energy distribution condition of the image, in this paper, we proposed a new image coding compression method which is based on logarithmic entropy as cost function to search the node labeled from bottom to top. Mother node log energy entropy was compared with son node and the smallest entropy node was marked until root node. The marked node constituted the wavelet packet tree and then the structure would be the optimal wavelet packet basis. The simulation results demonstrated that this method not only improved the compression ratio, but also promoted the effects of the reconstruction image.