基于编码通道数的JPEG2000压缩率控制算法

提出一种基于小波量化权值和编码通道数的JPEG2 0 0 0压缩率控制算法 .在位平面编码之前 ,根据当前码块所处的子带和码块内包含的编码通道数预先分配压缩码字数 .在编码过程中 ,编码码字数超过分配码字数时就截断码流 ,即实时截断码流 .与标准的率失真 (PCRD)优化算法相比 ,该算法不会在计算码率失真斜率和搜索最优分层截断点时产生运算开销 .实验结果表明 ,在 0 5b/像素下 ,所提算法对于Lena图像的存储器损耗和编码时间也只有标准PCRD优化算法的 6 8%和 18 7% ,但图像质量只低 0 8dB .这种算法资源损耗少 ,显著提高了运算速度 ,更有利于超大规模集成电路实现

Novel Scheme for JPEG2000 Rate Control Based on Number of Coding Passes

A novel scheme for JPEG2000 rate control was proposed based on the weight of wavelet quantification and the number of coding passes. The bytes of the compressed code steam were pre-assigned before coding the bit plane in terms of the sub-band and the number of coding passes contained in current code block. In the coding process, if the coded bytes greater than/equal to the bytes pre-assigned the code stream was truncated, i.e. truncating it in real time. Compared with the standard post-coding rate distortion (PCRD) algorithm, the proposed algorithm has no spending in calculating rate-distortion slope and searching for optimum truncation points. Code stream truncation could be fulfilled adaptively in code-block coding. Experiments show that for Lena image under 0.5 bit per pixel compression, the memory capacity needed and encoding time are only 6.8% and 18.7% of PCRD respectively, and the difference of the image quality between them is less than 0.8 dB. The scheme consumes less resource and greatly increases the encoding speed, so it is also better fitted to VLSI implementation.