Wednesday, September 30, 2020 Home   | Search
Viewing Abstract

Page: 443-454 ------------------------------> Last_modified :7/11/2017 12:11:00 PM

Title: Optimal Multilevel Image Thresholding: An Analysis with PSO and BFO Algorithms
Authors: V. Rajinikanth; N. Sri Madhava Raja; K. Latha
DOI:
Aff: St. Joseph’s College of Engineering, Department of Electronics and Instrumentation Engineering, Chennai 600 119, Tamilnadu, India.
Author Email:
Keywords: Otsu; Image Processing; Multi-Level Threshold; PSNR; SSIM
URLs: ABSTRACT-HTML  | FULLTEXT-PDF  | 
Abstract:Multilevel thresholding is widely adopted in image processing and pattern recognition fields. In this paper, Otsu based bi–level and multi-level image segmentation problem is addressed using Particle Swarm Optimization (PSO) and Bacterial Foraging Optimization (BFO) algorithms. Optimal thresholds are attained by analyzing histogram of the test image. Maximization of Otsu’s between class variance function is adopted to guide the heuristic algorithm based exploration. Performance of the proposed method is tested on eight benchmark test images using various numbers of thresholds. An assessment between PSO (constant weight), PSO (varying weight), Adaptive BFO, and Enhanced BFO are performed and the experimental results are validated using well known statistical parameters. For a bi-level optimization problem, considered heuristic algorithms show equal performance. For increase in threshold levels, PSO (constant weight) offers faster convergence and Enhanced BFO provides better structural similarity (SSIM) index.t