Constant variance enhancement: a digital processing technique

Abstract

An image enhancement process is one in which the image is manipulated for the purpose of increasing the information extractable by the human visual system. Constant variance enhancement is a technique which employs a high pass filtering to reduce the local average to zero for all regions of the picture and then applies a gain factor equal to the reciprocal of the local standard deviation to produce an output picture in which all local regions have equal variance. An example of a constant variance enhancement is presented with pictures illustrating the various mathematical stages of the processing operations. The concept of differential luminance gain is introduced as a crude measure of enhancement effectiveness and is numerically evaluated for the illustrative processing. The significance of negative differential luminance gain is considered, and ways of dealing with these spurious results are considered.