A Visual System Model and a New Distortion Measure in the Context of Image Processing.

摘要

A new error criterion for image processing is derived from a model of the visual system as a bank of narrowband spatial filters. It is a Fourier-amplitude weighted squared error criterion reflecting the change in probability of detection and channel phases under a variation in image intensity. This new error criterion is compared to the common squared intensity error criterion in optimal linear filtering of images in additive noise. The restorations with the new criterion are more visually pleasing,more sharply detailed,and truer in contrast than those for ordinary squared error,and impressive in their own right. The mathematical properties of the new criterion suggest that significantly increased efficiency in the storage or communication of images may be gained by its use.