基于上下文的四维矩阵变长编码  

Context-Based Four-Dimensional Matrix Variable Length Coding

作　　者：马行[1] 陈贺新[1] 赵岩[1] 齐丽凤[1] 

机构地区：[1]吉林大学通信新技术重点实验室,长春130012

出　　处：《吉林大学学报（信息科学版）》2007年第3期329-329,共1页Journal of Jilin University（Information Science Edition）

基　　金：国家自然科学基金资助项目（60372060）;科技部国际合作基金资助项目（2005DFA10300）

摘　　要：为了进一步提高彩色视频的压缩比，根据四维矩阵理论提出了基于上下文的四维矩阵变长编码方法。首先将彩色视频序列构建在一个统一的四维模型中，然后进行四维矩阵形式的预测、变换和量化。研究发现，此时各四维予矩阵系数间仍然存在一定的相关性，为了进一步利用这种相关性，提出了以四维矩阵视频予单元为整体，进行基于上下文的四维矩阵变长编码的方法，一方面可以节省对描述信息的表达，另一方面可以更加有效地利用上下文之间的相关性。实验结果表明，在保持高信噪比的情况下，该方法可以大幅度地提高压缩比。In order to increase the compression ratio of color video, a new coding method context-based four-dimensional matrix variable length coding was proposed according to four-dimensional matrix theory. The color video sequences were represented in mathematic model of four-dimensional matrix, and were predicted, transformed and quantized by four-dimensional matrix. In order to minimize the correlations between coefficients, the contextbased four-dimensional matrix variable length coding was proposed, which regarded the four-dimensional matrix video cell as a unit. It can save the expense of the describing information and take a good advantage of the correlations between the coefficients. Experimental results proved that the method can get twice times compression ratio than the advanced 3 dimensional digital cosine transform coding, and 15.05% of compression ratio increased than Huffman at high peak signal to noise ratio.

关 键 词：彩色视频 四维矩阵 上下文 变长编码 

分 类 号：TN919.8[电子电信—通信与信息系统]