摘要
本文提出了基于仿射变换的改进型矢量量化编码算法，并给出了两种不同的实用结构，与传统矢量量化算法相比，该方法在不需要重新训练新码本及不增加码本存储空间的情况下，降低了编码误差，使得重建图像的ＰＳＮＲ显著增加，图像的主观质量也得到很大的改善。
In this paper, an improved vector quantization (VQ) algorithm based on affine transformation is proposed for image compression. In addtion, two different new VQ structures are investigated in detail for practical applications. The proposed methods can significantly improve the performance of conventional VQ in terms of PSNR and visual quality without the need to generate an enlarged or specific codebook. Thus, the memory requirement is not increased, and a relatively small codebook size can also achieve fairly good reconstructed image quality.
出处
《通信学报》
EI
CSCD
北大核心
1998年第11期76-81,共6页
Journal on Communications
基金
国防科工委预研基金
国家自然科学基金
广东省自然科学基金
关键词
矢量量化
仿射变换
图像编码
Vector Quantization, Affine Transformation, Image Compression
