A context based adaptive arithmetic coding technique for lossless image compression

Significant progress has recently been made in lossless image compression using discrete wavelet transforms. The overall performance of these schemes may be further improved by properly designing efficient entropy coders. A new technique is introduced for the implementation of context based adaptive arithmetic entropy coding. This technique is based on the prediction of the value of the current transform coefficient, using a weighted least squares method, in order to achieve appropriate context selection for arithmetic coding. Experimental results illustrate and evaluate the performance of the proposed technique.