Modified Eight-Directional Canny for Robust Edge Detection

This article presents an edge detection mechanism, named as Modified Eight-Directional Canny for Robust Edge Detection (MEDC) which is a modification of Canny´s method. This technique has an add-in feature of performing edge detection in eight directions, by partitioning the two dimensional image space into eight planes in the order of 22.5⁰. This feature enables MEDC to detect almost all the edges of a given image. This method was tested on many textured-images and was found that it has produced better results. The performance of MEDC was analyzed qualitatively using the visual outputs as well as quantitatively in terms of Equivalent Number of Looks (ENL) and Edge Save Index (ESI). It is proven that MEDC performs better than Canny, on most of the test images. Hence, it may find a place in the process of image segmentation.