IMAGE REPRESENTATION USING NONORTHOGONAL BASIS IMAGES WITH ADAPTIVE WEIGHT OPTIMIZATION

Image-coding techniques that employ subsets of nonorthogonal basis ima ges chosen from two or more transform domains have been shown consiste ntly to yield higher image quality than those based on one transform f or a fixed compression ratio. However, these techniques have not been widely employed due to the very high computational complexity of exist ing formulations, This paper presents a new adaptive parallel mixed tr ansform technique for image representation, Formulations are derived, adaptive algorithms are described, and results of simulations are pres ented. Excellent performance is achieved with greatly reduced computat ional complexity compared to existing mixed-transform techniques, maki ng mixed-transform-based image coding practical.