图像的空间可分级压缩编码及其码率分配

Spatial Scalable Coding of Images and Optimal Bit Allocation

为了能使图像压缩编码算法同时具有较高的压缩率和良好的压缩性能,分析了SPIHT算法,在相同压缩效率的情况下,将其进一步改进,使之具有空间可分级性。在编码过程中,单个像素的相关信息按其所在的分辨率层输出到基础层或增强层码流中,而对于集合的测试信息按其最高层像素所在的分辨率层以较大的概率输出到相应的码流中。这样,虽然低频子带与高频子带相关联而编码,但实验结果显示与低频子带单独编码的性能非常接近。同时给出了一种基础层和增强层间的码率分配方案。实际结果显示:随码率的增大,基础层所占的比率越来越低;对包含边缘较多的图像,其基础层码率所占的比率也较低。

Image compression algorithms such as EZW and SPHT have a high compression rate. In order to they also have spatial scalability, SPIHT algorithm is analyzed. At the same compression rate, it is improved such that it has spatial scalability. In the proposed algorithm,the coding information of a isolated pixel （significant or insignificant） is outputted to the bitstream which the pixel is located;the coding information of a set is outputted to the bitstream which according to the highest layer pixels of the set is located. Although the low-frequency subband is encoded together with the high-frequency subbands,the experimental result show that the bit rate is very close to the results of the low-frequency subband is encoded independently. Meanwhile, an optimal bit allocation scheme between base-layer bitstream and enhanced-layer bitstream is also given in this paper. The simulation results show that the number of bit of base-layer bitstream is different from compressed images and the compression rate.