2-step wavelet-based edge detection using gabor and cauchy directional wavelets

Abstract

In this paper we propose a new wavelet-based approach for solving edge detection problem. Using primary knowledge of the direction of the edges extracted by ordinary separable and isotropic wavelet transform, we can classify the pixels as edge points, background and candidate edge point. For the candidate edge points, we propose implementing the Gabor and Cauchy directional wavelets and detecting the edge points precisely