Pattern matching image compression:: Algorithmic and empirical results

We propose a nontransform image compression scheme based on approximate one-dimensional pattern matching that we name Pattern Matching image Compression (PMIC). The main idea behind it is a lossy extension of the Lempel-Ziv data compression scheme in which one searches for the longest prefix of an uncompressed image that approximately occurs in the already processed image (e.g., in the sense of the Hamming distance or, alternatively, of the square error distortion). This main algorithm is enhanced with several new features 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 D, and so forth. These enhancements are crucial to the overall quality of our scheme and their efficient implementation leads to algorithmic issues of interest in their own right. Both algorithmic and experimental results are presented. Our scheme turns out to be competitive with JPEG and wavelet compression for good quality graphical images. We also review related theoretical results.