A Novel Statistical Thresholding in Edge Detection Using Laplacian Pyramid and Directional Filter Banks

摘要

To achieve good localization and good detection of edges, a multi scale approach is required along with the direction of location of the edge. In this paper an edge detector is proposed which uses the Laplacian Pyramid for multi scale decomposition and a Directional Filter Bank to extract the direction. Furthermore, a thresholding based on statistical principles is used to validate the eligibility of a pixel to be an edge pixel. The results obtained from the proposed method are compared with the most popular Canny edge detector and are found to be appreciable. The results are also compared in terms of quantitative terms like psnr (dB) and Mean Square Error (MSE). The proposed method is found to performing better than the Canny edge detector both in terms of quantitative and qualitative evaluation.