Lossless bit-plane compression of images with context tree modeling

Bit-plane coding has been widely used in lossless compression of gray-scale images or color palette images. In the context-based encoding procedure, the values of previous significant planes are also considered, which is simplified as surrounding pixel´s expectation values. In this paper, we further improved the encoding scheme by context-tree modeling, in which both the order and depth of the context template is optimized. In probability estimation scheme, a forgetting factor and context weighting are utilized. According to our experiments, the proposed method compete with the existing algorithms in lossless compression of gray-scale images, and outperforms all competitive methods in compression or color palette images.