关键词: 信息处理技术, 三维矩阵, 彩色图像压缩, 可变分割, 标量量化

Abstract: Fixed size matrix segmentation and fixed size transform are normally used in the color image compression methods based on three dimensional matrix wide discrete cosine transform (3 DMWDCT). However this fixed size segmentation can not effectively exploit image relativity since different areas of an image have different statistical properties. Here a color image compression algorithm using variable matrix size 3 DMWDCT and based on variable sized matrix segmentation was proposed. This algorithm divides the original color image into variable sized subthreedimensional matrices according to the activity characteristics of the image. Then the variable sized 3 DMWDCT was applied to the corresponding sized subthreedimensional matrices. DCT coefficients were quantized through nonlinear scalar quantization and entropy coding. Experimental results show that the compression efficiency of the proposed algorithm is better than that of JPEG at low bit rate. Compared with JPEG, the proposed algorithm improves PSNR at most 2dB, and the quality of the reconstructed images is improved in subject evaluation.

Key words: information processing, threedimensional matrix, color image compression, variablesized segmentation, scalar qantization