A context based arithmetic image coding using a separate least squares algorithm

The overall performance of discrete wavelet transforms for lossless image compression may be further improved by properly designing efficient entropy coders. In this paper a novel technique is proposed for the implementation of context-based adaptive arithmetic entropy coding. It is based on the prediction of the value of the current transform coefficient. The proposed algorithm employs a weighted least squares method applied separately for the HH, HL and LH bands of each level of the multiresolution structure, in order to achieve appropriate context selection for arithmetic coding. Experimental results illustrate and evaluate the performance of the proposed technique for lossless image compression.