| Autor: |
Lan, Lin, Wang, Shengsheng |
| Předmět: |
|
| Zdroj: |
Multimedia Tools & Applications; May2024, Vol. 83 Issue 15, p45241-45290, 50p |
| Abstrakt: |
Multi-level threshold image segmentation is widely used in medical image segmentation. Traditional methods for selecting optimal thresholds suffer from exponentially increasing time complexity as the number of threshold levels increases. In order to solve these problems, we choose African vultures optimization algorithm (AVOA) and introduce a novel modified African vultures optimization algorithm method called OLAVOA that combines predatory memory and logarithmic spiral based on opposition learning. In addition, we also apply 2D Kapur's entropy as a fitness value function of OLAVOA for multi-threshold image segmentation to solve the problem of traditional method. In the 30 benchmark function experiments at IEEE CEC2014, the average value of the experimental results of OLAVOA mostly outperforms the other algorithms. In the experimental convergence graph, OLAVOA's performance showed its convergence's velocity and capacity to depart from the local best. In addition, to demonstrate the effectiveness of OLAVOA on medical image segmentation, the image segmentation experiment included chest X-ray image of patients with COVID-19 and brain MRI image. Additionally, it was demonstrated that OLAVOA outperformed other approaches in segmentation trials by having greater adaptability in different threshold levels. Therefore, OLAVOA is effectively utilized to segment medical images. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
Full text from SpringerLink