Corner Detection Based on Scalable Masks

Abstract

Scalable masks for the selection of angular structures in two-dimensional (2D) digital images are considered. The mask is a 2D window sliding over the image and convolved with image fragments. We propose the model of a scalable 2D mask based on expanding smaller mask along its sides and edges. In this case, the submatrices remain unchanged, and the generation of new elements consists in repeating the elements of the submatrix, preserving the structure of the corner. Numerical experiments with test images are performed.