ABSTRACT
Edges and corners contain signature attributes of describable shapes. Thus, many target recognition algorithms rely on the detection of these features. One approach of detecting edges of any orientation, clarity, and length is to analyze the correlation responses of the image with a set of Gabor filters. This chapter reviews the construction of these filters and presents a few examples of implementation. Gabor filters are a set of correlation filters designed to respond to edges of various orientations, scale, and length. A properly designed suite of Gabor filters can respond to all edges in the image or set of images. A correlation with a Gabor filter is equivalent to isolating a small set of frequencies, which is related to the distance from the DC term. These frequencies are only for a small rotation angle, which is related to the rotation of the spots in the image.
