Image processing for quality improvement

Abstract

In order to encourage the development of computer-based methods which produce better quality pictures, the long and successful history of image processing in photography and graphic arts is called to the attention of the computer fraternity. Studies of contrast sensitivity and of the relationships among illumination, object reflectance, and image luminance are presented. Homomorphic filtering is analyzed in this light. Adaptive filtering methods which exploit perceptual phenomena as well as the physical properties of imaging systems are given. These methods attempt to permit the attainment, of a very high degree of sharpening, equally visible throughout the tone scale as well as in image areas of quite different character, without the concomitant appearance of quality-destroying artifacts such as haloes or overshoots. Results are compared with linear and homomorphic nonadaptive sharpening.