An Optimized Design of Bit Plane Prediction Based Image Compression Algorithm on DSP
Zheng Zhao, Hongxu Jiang, Bo Hu Li · 2009
According to the characteristics of DSP platform, this paper presents an optimized design of wavelet image compression algorithm based on bit plane prediction. The optimization is made focus on adaptive quantization module and predictive and entropy coding module respectively. As for the adaptive quantization module, the optimization is made through the ways of luminance value calculation by look-up table, improving calculation formula of standard deviation, eliminating storage space of luminance values under the condition that no time complexity is increased, sampling calculation of standard deviation and converting heavy division calculation to multiplication calculation. As for the predictive and entropy coding module, we improve the storage method of the wavelet coefficients, implement predictive classification by look-up table, and optimize the scanning encoding method by zero block coding. The tested results through hardware platform on DSP show that the operation speed of the optimized algorithm has increased by about 30 times and reaches more than 70 Mbps.