On the impact of histogram sparseness on some lossless image compression techniques

Most of the image compression techniques currently available were designed mainly with the aim of compressing continuous-tone natural images. However, if this assumption is not verified, such as in the case of histogram sparseness, a degradation in compression performance may occur. In this paper, we analyze the impact of histogram sparseness in three state-of-the-art lossless image compression techniques: JPEG-LS, CALIC and lossless JPEG-2000. Moreover, we propose a simple procedure for on-line histogram packing, which holds nearly the same improvement as offline histogram packing. Results of its effectiveness when associated with JPEG-LS are presented.