论文部分内容阅读
在彩色图像的三维离散余弦变换压缩体系中,交流系数游程编码后存在大量的长零游程,传统的JPEG基于(Run,Level)统计进行熵编码,压缩性能较差;通过对交流系数分布的研究,提出基于(Run Level,Level)的联合概率分布进行熵编码的改进方法,该方法码表简单,且可直接复用JPEG码表,有很大潜力应用于更高维的视频图像变换压缩领域。另外,根据非零系数的分布特点和编码效率分析,对密集区和疏散区域采用不同的编码方法。实验结果表明,在相同的PSNR下,较JEPG基线标准有11%的码流节省。
In the three-dimensional discrete cosine transform and compression system of color image, there are a lot of long-zero runs after the run-length coding of AC coefficient. The traditional JPEG performs entropy coding based on (Run, Level) statistics and has poor compression performance. By studying the distribution of AC coefficients , An improved method of entropy coding based on (Run Level, Level) joint probability distribution is proposed. The method has the advantages of simple code table and direct reuse of JPEG code table, and has great potential in the field of higher-dimensional video image transform and compression . In addition, according to the distribution characteristics of non-zero coefficients and coding efficiency analysis, the dense area and evacuation area using different coding methods. Experimental results show that there is 11% code stream savings over the JEPG baseline standard for the same PSNR.