Lossless Image Compression by PPM-Based Prediction Coding

Abstract

Most of speech and image data compressed by lossy compression whose decompressed data are different from the original ones. Here, the different between the decompressed data and the original ones cannot be recognized by most of people. Lossless image compression, which gives exactly the same decompressed data as the original ones, is necessary for medical image, art work image, and satellite image, which are frequently processed by computers now. This paper proposes lossless image compression by prediction coding whose frequency table operation is based on PPM(Prediction by Partial Match). The prediction algorithm for the proposed method is based on that for CALIC, an existing lossless image compression method; and the difference between the predicted value and the actual one is encoded by PPM-based compression method. In this compression method, initial values in the frequency table and frequency table operation method are modified to achieve efficient compression ratio. Computer simulation says that the compression ratio of the proposed method is better than that of CALIC by about 0.07 bit/pixel.