,
Annamalai University , Chidambaram , India
,
New Horizon College of Engineering , Bengaluru , India
,
Kangeyam Institute of Technology , Tirupur , India
,
Kangeyam Institute of Technology , Tirupur , India
,
St. Joseph’s Institute of Technology , Chennai , India
St.Joseph's College of Engineering , Chennai , India
Image segmentation plays an important role in medical diagnosis and recognition, but the traditional methods of multilevel thresholding have exponential computation complexity with the number of thresholds. The study corresponds to the necessity to have a computationally effective parameter-free optimization to support fast clinical decision-making. The study suggests two optimization systems that are used to optimize image segmentation, namely the Jaya algorithm and Stochastic Fractal Search (SFS). Jaya algorithm, with its single-phase update mechanism and no algorithm-specific parameters, is used to calculate optimal thresholds based on the maximization of Entropy in Kapur. At the same time, the SFS algorithm is based on the idea of natural fractal patterns and the diffusion of particles, which are used to maximize the between-class variance of Otsu. These two techniques were strictly tested on 256 × 256 8-bit benchmark images (Cameraman, Lena, and Peppers). The results of numerical assessments indicate that the two algorithms are able to approach optimal threshold values irrespective of varying levels (K = 2, 3, 4, 5). In the Jaya algorithm, especially, the computational efficiency was much better, and the minimum processing time was used without compromising the quality of segmentation. When compared to the existing metaheuristics such as GA and PSO, it is shown that the suggested methods are more stable and robust and do not change in levels even when the number of thresholds grows. The results place the Jaya and SFS algorithms as promising algorithms to perform multi-level thresholding of images. They are very appropriate in real-time medical imaging and other technical applications that demand high-quality segments with minimal computational overhead.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
0
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.