Abstract
Accurate segmentation is vital for improving diagnostic precision and supporting various medical applications. However, traditional methods like Kapur's and Otsu's, while effective for bi-level thresholding, face challenges when applied to multi-thresholding due to their high computational demands. To address these limitations, we introduce a novel multi-level threshold segmentation method for chest X-ray images, utilizing an Improved Multi-Objective African Vultures Optimization Algorithm (IMMOAVOA). This algorithm integrates an average partial opposite learning strategy with an in-depth exploration mechanism, enhancing the balance between exploration and exploitation. Additionally, a multi-objective thresholding model that combines Otsu's method with two-dimensional (2D) Kapur's entropy is developed to leverage the strengths of various thresholding techniques.The performance of the proposed method is evaluated using ZDT and DTLZ test functions, as well as chest X-ray images. Experimental results demonstrate that IMMOAVOA outperforms the original AVOA and other benchmark algorithms across all evaluation metrics, proving its superior efficiency in multi-threshold image segmentation.