Fuzzy control of subband coded image quality using standard and fuzzy quality measure

We propose a method for subband coded image quality supervision using a clo sed loop fuzzy controller for adjusting the quantization steps. As a measur e of the image quality used in the control loop, we first use the standard quadratic quality measure SNR in decibels. We then expand this measure usin g the compression ratio in a novel quality measure that is designed on the basis of subjective experiments and implemented with fuzzy logic. Our resul ts demonstrate the superior performance of the proposed fuzzy quality measu re over the standard SNR quality measure in terms of correlation with the s ubjective impression of compressed images. The proposed method should prove to be useful, especially for natural images that have decaying power spect ral density. (C) 2001 Society of Photo-Optical Instrumentation Engineers.