A New Image Pre-Processing for Improved Performance of Entropy Coding

A new pre-processing method through a newly proposed image transform scheme is presented in this paper. It enhances the efficiency of entropy encoders in the compression of images. Before an entropy encoder encodes a stream of pixel values in an image, the proposed method adaptively performs the new image transform under various transform patterns and then selects the transformed image with the least entropy as the pre-processing output. Finally it transmits the adjusted data to the entropy encoder. Statistical characteristic is enhanced after this pre-processing step. The experimental results using 8-bit gray-scale images illustrate that the proposed method efficiently improves the performance of entropy coding by reducing the compression ratio to a much lower level than that of plain entropy coding.