A proposal of edge detection in images with multiplicative noise using the Ant Colony System algorithm. (April 2022)

• Record Type: Journal Article
• Authors: Baltierra, Sergio; Valdebenito, Jonathan; Mora, Marco
• Is Part Of: Engineering applications of artificial intelligence. Volume 110(2022)
• Publication Date: 2022-04
• Subjects: Edge detection -- Ant Colony System -- Multiplicative noise -- Coefficient of variation; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2022.104715
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Multiplicative noise is one of the most aggressive types of noise present in various types of images: Synthetic Aperture Radar, Ultrasound images, Ultrasonic Imaging, among others. Edges detectors such as Canny or Sobel are not very efficient for processing an image with multiplicative noise, they also require filtering algorithms, a preprocessing, which is why bio-inspired algorithms are an alternative for processing images with the presence of multiplicative noise, due to its efficiency in finding an approximate solution. This article proposes a method for the edges detection in images with multiplicative noise using the Ant Colony System algorithm. For which we must adapt the Ant Colony System algorithm to detect contours, this we define the calculation of a global pheromone matrix between several edge detection equations, gradient, and the coefficient of variation, these equations are compared for their edge detection performance using a visual inspection and a performance function. The results of the experiments show a correct implementation of the algorithm proposed to the images with multiplicative noise even for high noise levels. Highlights: Selection of an Ant Colony System algorithm representative of literature. Adaptation of the selected algorithm using a coefficient of variation. Robust edge detection in images with multiplicative noise. Objective evaluation of results concerning the ground of the truth.