Novel and improved methods of regular geometric shape recognition from digital image using artificial ants

Abstract

This contribution demonstrates how artificial ants can extract regular geometric shapes from grey scale images. We propose here two methods the first of which is a modified version of existing Ant System algorithm. The second method proposed is Ant Regeneration and Recombination System (ARRS). Our schemes comprise of three steps. Firstly, MATLAB edge detection operator converts a grey scale image into a binary one. Our schemes are then applied on this binary image to detect closed loops. Finally, these closed loops are tested for different geometric shapes. The schemes with incredible time and memory efficiency can detect both intersecting and non intersecting regular shapes.