SEGMENTATION AND CLASSIFICATION OF EDGES USING MINIMUM DESCRIPTION LENGTH APPROXIMATION AND COMPLEMENTARY JUNCTION CUES

This article presents a method for segmenting and classifying edges us ing minimum description length (MDL) approximation with automatically generated break points. A scheme is proposed where junction candidates are first detected in a multiscale preprocessing step, which generate s junction candidates with associated regions of interest. These junct ion features are matched to edges based on spatial coincidence, For ea ch matched pair, a tentative break point is introduced at the edge poi nt closest to the junction. Finally, these feature combinations serve as input for an MDL approximation method which tests the validity of t he break point hypotheses and classifies the resulting edge segments a s either ''straight'' or ''curved.'' Experiments on real world image d ata demonstrate the viability of the approach. (C) 1997 Academic Press .