Fuzzy logic and the principle of least commitment in computer vision

Abstract

The application of fuzzy logic to computer vision processes has grown rapidly. The appeal of such techniques stems, at least partially, from the fact that they naturally maintain multiple hypotheses to varying degrees until a crisp decision must be made. This satisfies the principle of least commitment, originally stated by David Man for the development of intelligent computer vision algorithms. In this paper, we demonstrate fuzzy set theory's support of this principle with three examples from midlevel vision.