Efficient Entropy Coding for Image Compression

In this paper we propose entropy coding algorithms for image and video compression systems using binary decomposition. We show that this technique allows for efficient coding algorithms. We propose two decompositions that allow for reduction of the number of coding parameters by a factor of about 10 compared to the JPEG, while the compression performance is not worse. A small number of coding parameters per context (3-4) may be beneficial for multi-context algorithms and for systems with limited memory or hardware resources, e.g., for mobile communication devices.