基于改进进化规划的图像压缩中的矢量量化

Vector Quantization in Image Compression Based on Improvement Evolutionary Programming

提出了基于改进进化规划和LBG算法的一种新的矢量量化算法。由于进化规划中十进制编码方案的收敛速度比二进制编码方案快,因此在进化规划中采用十进制编码方案。同时,根据图像压缩的特点,为了提高算法性能,该文采用一种自适应的多位变异操作。之后将改进的进化规划引入到图像压缩的矢量量化问题中。实验表明,新算法在提高恢复图像质量方面比传统方法优越。

A new vector quantization algorithm for image compression is presented. It is based on improvement evolutionary programming and LBG algorithm. Firstly, this paper proposes a new evolutionary programming algorithm. Since the convergence speed in decimal coding is quicker than that of binary coding, it is reasonable that novel evolutionary programming employs decimal coding with several mumation bits to improve the performance according to the characteristic of the image compression. Secondly, vector quantization algorithm for image compression is done. The experiments show that this new algorithm performs better than the conventional schemes in the aspects of high image quality.