基于上下文的四维矩阵变长编码

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。将彩色视频序列构建在一个统一的四维模型中,进行四维矩阵形式的预测、变换和量化。为了进一步利用系数间存在的相关性,对四维矩阵视频子单元进行基于上下文的四维矩阵变长编码,从而节省了对描述信息的表达,并且更加有效地利用了上下文之间的相关性。实验结果表明,基于上下文的四维矩阵变长编码方法在保持高信噪比的情况下,与改进的三维离散余弦变换算法相比,压缩比可以提高1倍;与霍夫曼编码方法相比,压缩比可以提高15.05%。

Context-Based Four-Dimensional Matrix Variable Length Coding

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 context-based 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.