Improving VQ Index Compression with Modified Relative Index Table

Image Vector Quantization (VQ) has many current applications, like speech and image compression, and envisioned applications, such as digital watermarking, data hiding and speaker identification. Given an image, VQ produce an index map by quantizing the image block by block. Recently, a loss less VQ index compression scheme, called Index Searching Algorithm with index Associated List (ISAIAL), was proposed to improve the VQ performance by using relative index table. The basic idea behind ISAIAL is to utilize the effect of correlation between adjacent indices in the index map. In this paper, two modified relative index tables are proposed to improve the performance of ISAIAL further. The computation of optimal size for modified relative index tables is also addressed. The proposed methods were evaluated via extensive experiments. The results show that the ISAIAL is improved by the proposed methods. Moreover, our approaches apparently reduce the bit rate as compared to the conventional VQ and some commercial loss less compressors, such as rar, zip and arj.