Evaluation of Some Reordering Techniques for Image VQ Index Compression

Frequently, it is observed that the sequence of indexes generated by a vector quantizer (VQ) contains a high degree of correlation, and, therefore, can be further compressed using lossless data compression techniques. In this paper, we address the problem of codebook reordering regarding the compression of the image of VQ indexes by general purpose lossless image coding methods, such as JPEG-LS or CALIC. We present experimental results showing that techniques available for palette reordering of color-indexed images can also be used successfully for improving the lossless compression of images of VQ indexes.

• Type: Conference
• Year: 2004
• Where: ICIAR
• Authors: António R. C. Paiva, Armando J. Pinho
• Keywords: ICIAR 2004, Lossless Data Compression, Purpose Lossless Image, VQ Indexes