Custom Dilated Edge Detection Filters

Abstract

Edge detection is a basic and fundamental function in image processing. Inspired by the new dilated convolution techniques which have impressive results in machine learning, we discuss here the idea of dilating the standard filters which are used to compute the gradient of an image. We compare the result of the proposed dilated filters with original filters and extend or custom variants of them. We also present the impact of dilatation on a complex edge detection algorithm, like Canny.