Gray scale image compression based on multiple-valued input binary functions, Walsh and Reed-Muller spectra

Abstract

A new method for the lossless compression of gray scale images has been proposed. Coding of intensities is first applied to make the data more amenable for compression. A prediction process is performed followed by the mapping of prediction residuals. The prediction residuals are then split into bit planes to which the compression technique is applied. These bit planes can be coded as uncompressed, expressed as minterms or compressed using a variable block-size segmentation and coding. A dictionary of patterns is formed from simple multiple-valued input binary functions, basic Walsh, triangular Reed-Muller weights and some frequently occurring patterns. Other compression methods used in our scheme include minterm coding, coordinate data coding, Generalized k-Variable Mixed-Polarity Reed-Muller expansion and the reference row technique. The proposed scheme has been implemented in the C language and compared with other stare-of-the-art techniques.