A Practical Wavelet Compression for Arbitrarily-Sized Natural Color Images

Abstract

A practical method for wavelet-based image compression was proposed, where ‘practical’ means that this method can be performed on typical images of any size, whether grey-scaled or color images. Its performance with simple methods used for wavelet decomposition of an arbitrarily-sized image were compared . Color images were transformed into the appropriate domain before the wavelet decomposition was employed in the same way as grey-scaled images. The wavelet coefficient matrices were encoded with an existing bit-plane encoding algorithm of the authors, (non-list context adaptive wavelet difference reduction). With its fast and memory-effective algorithm, the proposed coder performed at the same level as the JPEG2000 standard and significantly outperformed the existing JPEG standard for all test images which were downloaded from the Internet.