Contour Detection in Noisy Pictures Using Heuristic Search Methods.
NEW YORK UNIV BRONX DEPT OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCE
Pagination or Media Count:
A method for detecting contours in noisy pictures is presented. The problem of contour detection is reduced to the problem of finding a minimal-cost path in a weighted graph. A figure of merit takes into account the properties of the contour, and graph searching techniques are used to find the contours which minimize the given figure of merit. Because of the global approach, this method has great flexibility. Some experimental results are given, which show how information about the shape of the object to be detected can be incorporated in the figure of merit, thus allowing the extraction of contours from noisy pictures and the separation of touching objects. Author
- Computer Programming and Software
- Computer Hardware