The Comparison of the Effects of Thresholding Methods on Segmentation Using the Moth Flame Optimization Algorithm
Year: 2023
Type: article
Abstract: Segmentation is an important preprocessing step that directly affects the success in image processing applications. There are many methods and approaches used for the segmentation process. Thresholding is a frequently used approach among these methods. There are several suggested approaches to thresholding. In this study, six different thresholding approaches were used as the fitness functions using the moth flame algorithm and the results obtained from these approaches were compared. In experimental studies, seven different threshold levels of 10 different images were studied. In comparisons made with three different metrics, it was seen that the Otsu method was generally more successful. It has also been observed that the minimum cross entropy and Renyi entropies can be used as alternatives.
Author: Murat KARAKOYUN
Institution: Necmettin Erbakan University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4379196573