Computer Engineering and Applications ›› 2013, Vol. 49 ›› Issue (14): 147-151.

Previous Articles     Next Articles

Mean-value and vector quantization hybrid algorithm for high speed image coding

WANG Dongfang1, LI Jia2, YU Ningmei1   

  1. 1.Department of Electronics Engineering, Xi’an University of Technology, Xi’an 710048, China
    2.Department of Electronics and Mechanics Engineering, Xi’an Vocational and Technical College, Xi’an 710032, China
  • Online:2013-07-15 Published:2013-07-31

均值与矢量量化复合快速图像编码算法

王冬芳1,李  佳2,余宁梅1   

  1. 1.西安理工大学 电子工程系,西安 710048
    2.西安职业技术学院 机电工程系,西安 710032

Abstract: As an effective technology for data compression, Vector Quantization(VQ) is widely used in the field of data coding because of its simple algorithm and high compression rate. According to analyzing the character of image block, a hybrid algorithm combining mean-value and VQ is presented in this paper, which encodes smooth image block and nonsmooth block by mean-value and VQ algorithm respectively. The presented algorithm can save the storage space by removing the smooth codewords in the codebook and speed up encoding process. In order to improve the storage efficiency, the algorithm which can compress original codebook to 1/8 by rotating and reversing codewords is adopted. At the same time, the search algorithm is optimized using Extend Block Nearest Neighbor(EBNNS) method. The whole system encoding speed can achieve about 7.7 times faster than full search VQ algorithm.

Key words: image coding, vector quantization, mean-value coding

摘要: 矢量量化是一种有效的数据压缩技术,由于其算法简单,具有较高的压缩率,因而被广泛应用于数据压缩编码领域。通过对图像块灰度特征的研究,根据图像的平滑与否,提出了对图像进行均值和矢量量化复合编码算法,该算法对平滑图像块采用均值编码,对非平滑块采用矢量量化编码。这不仅节省了平滑码字的存储空间,提高了码书存储效率,并且编码速度大大提高。同时采用码字旋转反色(2R)压缩算法将码书的存储容量减少到1/8,并结合最近邻块扩展搜索算法(EBNNS)对搜索算法进行优化。在保证图像画质的前提下,整个系统的图像编码速度比全搜索的普通矢量量化平均提高约7.7倍。

关键词: 图像编码, 矢量量化, 均值编码