A New Method for Lossless Image Compression Using Recursive Crack Coding

Abstract

Popular entropy coding methods for lossless compression of images depend on probability models. They start by predicting the model of the data. The accuracy of this prediction determines the optimality of the compression. These methods are very slow because of visiting the data (pixels) in left to right order. Parallel implementation of these methods is adopted by researchers to speed up the process. In this paper, the authors propose a new approach to image compression using crack coding. The novelty and better compression ratio of the method is due to its recursiveness in finding the variable-length entropy. The proposed method starts with the original image and develop crack codes in a recursive manner, marking the pixels visited earlier and expanding the entropy in four directions. The proposed method is experimented with sample bitmap images and results are encouraging. The method is implemented in uni-processor machine using C language source code.