Practical use of the rebound effect in image processing

This paper discusses the use of the rebound effect in image processing. It shows that the method is effective for various problems: the correction of geometrical noise, selection, tracking, contour discrimination, and combining observation channels. This approach depends only on the properties of the input image. The optimum regime for a specific image-processing problem is found completely automatically. In this case, no a priori: assumptions are made concerning the starting data, and no parameters are given. (C) 1999 The Optical Society of America. [S1070-9762(99)01504-3].