Contents 1. Introduction
Image compression methods have become very important for communications and multimedia applications. Among lossy block image encoding methods, JPEG and vector quantization (VQ) [1], are famous and classical methods. VQ is proven to be an efficient and effective scheme, especially for image like computer graphics that has many abrupt discontinuities in spectrum domain. VQ uses a look up table (called codebook) principle by template matching in spatial domain so as to find a closest item (winner code) to the input image block within the table according to a certain distortion measure (usually Manhattan distance or Euclidean distance). Then VQ only transmits the index of winner code instead of winner code itself to reduce the amount of image data. Because an exactly the same table (codebook) is also available at the receiving end, the image can be decoded in inverse look-up table way by using received index of winner code and then winner code itself is pasted to corresponding position of the image to reconstruct it. Therefore, VQ features a rather heavy encoding process due to a lot of matching computation and a very simple decoding process.
