【摘要】 针对传统视频压缩中量化后的8×8 DCT块系数,提出了一种新的熵编码算法.该算法在H.264的自适应变长码编码(CAVLC)的基础上,把游程(run)和层次(level)分离,并采用规则码表进行编码,从而保证了码字可以通过简单计算获得,避免了传统的查表法,降低了视频编解码系统的存储要求.在编码时,该算法充分利用已有的编码信息,通过不断转换码表实现自适应性.算法根据8×8 DCT块中的0系数个数,分3种不同的方式编码run,从而提高了编码效率.实验结果表明,本算法在各种编码条件下都能实现很好的压缩效果.更多还原

【Abstract】 A novel algorithm for entropy coding was introduced to code block after discrete cosine transform and quantization in conventional video compression. The algorithm was based on context-adaptive variable length coding(CAVLC) of H.264.Level and run were coded thus separately using inerratic code tables.All code words could be obtained through a few simple calculations avoiding the look-up table,which substantially reduced random-access memory(RAM) requirement at encoder or decoder systems. The algorithm selected different code tables according to the known coding information,thus ensuring the context-adaptive performance.Three ways were adopted to code run according to the number of zero coefficients in block,which improved the compression efficiency.Simulation results show that the algorithm achieves excellent performance of compression in various coding conditions.更多还原