Edge feature analysis by a vectorized feature extractor for multiple edges

Abstract

In this paper we propose a method to generalize a one-dimensional edge feature extractor to two dimensions and clarify its properties for edges close to each other (i.e., multiple edges). Previously we stated the necessity of a measure that can discriminate a clear edge with small edge height from a noisy edge with large edge height. Then we proposed that the edge features be viewed as a composition of edge height and edge reliability, based on the analysis of variances within a window around the edge. From this analysis, we constructed an edge feature extractor in 1D. In this paper we generalize our 1D edge feature extractor to a 2D vectorized edge extractor and prove that it can calculate an edge orientation and edge height accurately and also can reduce computation time. Experiments demonstrate this clearly. The same vectorization technique is also applicable to the Canny operator. In multiple edges cases, we derive the conditions for calculating accurate edge locations. Since most edge detection methods use local maxima of the edge height function to detect edge points, we find the conditions by differentiating the edge height function. © 1998 Scripta Technica, Syst Comp Jpn, 28(10): 20–29, 1997