摘要： 将彩色图像的R、G、B三帧作为整体考虑，并利用三维离散余弦变换（3D-DCT）进行压缩编码是彩色图像编码的新方法。为了进一步提高基于该方法的彩色图像编码效率，该文提出了基于三维块特征分类的彩色图像编码算法。实验结果表明，该方法同传统彩色图像编码方法相比，在保证相同的恢复图像效果时，将压缩比提高了约22%，也能够进一步提高基于3D-DCT的彩色图像压缩编码的性能。

关键词: 三维块, 特征分类, 三维离散余弦变换, 彩色图像, 编码

Abstract: It is a new idea of color image coding algorithm that considers R, G, B frames of a color image as a whole and makes use of three-dimensional discrete cosine transform (3D-DCT) to compress. In order to improve the color image coding efficiency based on this method, the new color image coding algorithm based on three dimensional feature classification is proposed. The experimental results show that the proposed method can improve the compression ratio by about 22 percent with the same quality of reconstructed image compared to traditional color image coding method. It can also improve the performance of color image coding based on 3D-DCT.

Key words: three-dimensional block, feature classification, 3D-DCT, color image, coding

中图分类号: 

• TP391