Pattern Matching Image Compression with Predication Loop: Preliminary Experimental Results

Recently, a novel image compression technique based on pattern matching was proposed, namely Pattern Matching Image Compression (PMIC). Basically, it is a lossy extension of the well known Lempel-Ziv scheme. It was proved that such an extension leads to a suboptimal compression, and that the compression ratio can as low as the so called R enyi entropy. Success of PMIC crucially depends on several enhancements such as searching for reverse approximate matching, recognizing substrings in images that are additively shifted versions of each other, introducing a variable and adaptive maximum distortion level, and so forth. In this paper, we introduce another enhancement, namely, predictive coding. More importantly, we implement Di erential Predictive Code Modulation (DPCM) within PMIC scheme. We report here some preliminary experimental results which show that PMIC enhanced by prediction can improve compression ratio for good quality images as well as speed of compression for PMIC. This work was supported by NSF Grants NCR-9415491 and CCR-9201078.