论文部分内容阅读
近几年来,矢量量化在图象高效编码的应用研究中展示出值得注意的性能。本文探讨了变换域矢量量化的压缩机理,为有效地将变换技术与矢量量化相结合,基于Walsh-Hadamard变换(WHT)的特性提出WHT域乘积码矢量量化(WHT-PC-VQ)方案。WHT-PC-VQ的复杂度较低。模拟实验结果表明其性能优于其它非自适应图象编码方案。WHT域中的矢量量化比DCT域中的矢量量化更具潜力。这表明矢量量化与传统技术相结合所面临的新问题是矢量量化应用研究中的重要课题。
In recent years, vector quantization has shown notable performance in the application research of efficient coding of images. In this paper, we discuss the compression mechanism of vector quantization in transform domain. In order to combine the transform technique with vector quantization effectively, we propose a WHT-PC-VQ scheme based on the characteristics of Walsh-Hadamard transform (WHT). WHT-PC-VQ is less complex. Simulation results show that its performance is superior to other non-adaptive image coding schemes. Vector quantization in the WHT domain has more potential than vector quantization in the DCT domain. This shows that the new problem that vector quantization combines with traditional techniques is an important issue in vector quantization applications.