Edge detection by point classification of Canny filtered images

Abstract

Listen

Translate article

This paper presents a new strategy that exploits a neural classifier to select candidate edge points from a filtered image. First, a spatial filtering for edge enhancement (the Canny filter) is used to calculate a set of large variation points, corresponding to the local maxima of the filtered image. A preliminary coarse selection is then performed, which exploits neighbourhood information to produce an extended pseudo-edges set (PES). Finally, a features' vector is computed for each point belonging to the PES, and fed into a classifier that decides whether it belongs to the target edge set or not. Since the selection works at the PES level, the creation of data sets for the training and testing of the classifier was performed in a fast and easy way by means of a computer-aided interactive tool. Experimental results proved that the proposed selection criterion is effective in improving the performances of the detector over classical threshold methods (e.g., the hysteresis selection used by Canny).