Approximation quality improvement techniques in progressive image transmission

The present investigation is concerned with methods for improving the quality of the intermediate approximations in progressing image transmission. Interpolating schemes are examined, taking into account a quality measure of the interpolation results. Attention is given to the nonuniform development of the pyramid, and the progressive pixel quantization. A combined use of the three presented techniques is found to improve greatly intermediate approximations in progressively transmitted images, minimizing the amount of information needed to reach a certain objective and subjective quality level in the receiver. All the developed schemes are intrinsically independent of the selected coding algorithm used to generate the hierarchy on which the progressive transmission is based.