Brain MRI Segmentation Using an Improved Bat Optimization Algorithm

Magnetic Resonance Imaging (MRI) is a non-invasive and widely adopted diagnostic modality that provides high-resolution information for the assessment of neurological disorders and the characterization of brain structures. Accurate segmentation of brain MRI images is a fundamental step in computer-aided medical image analysis because it enables the identification and spatial delineation of anatomical regions and provides essential information for subsequent diagnostic and clinical decision-making processes. However, conventional manual segmentation performed by radiologists is time-consuming, labor-intensive, and subject to inter- and intra-observer variability. Moreover, the complex morphology and intensity variations of soft tissues, together with image noise and system-induced artifacts, can substantially reduce segmentation accuracy and make the reliable identification of tissue boundaries challenging. To address these limitations, this study proposes an automated brain MRI segmentation framework based on an Improved Bat Optimization Algorithm (IBOA). The proposed approach integrates the global search capability of the bat-inspired optimization mechanism with the K-means clustering algorithm, in which the optimization procedure is employed to determine more appropriate initial cluster centroids. This hybrid strategy aims to reduce the sensitivity of conventional K-means clustering to the random initialization of centroids and consequently improve the reliability and accuracy of the segmentation process. The performance of the proposed framework is evaluated through quantitative simulations implemented in the MATLAB environment and is benchmarked against alternative state-of-the-art segmentation approaches. The experimental results demonstrate that the proposed method achieves a segmentation accuracy of ۹۹.۵%, consistently outperforming the considered baseline methods. These findings indicate that the integration of metaheuristic optimization with clustering-based segmentation can provide an effective computational framework for accurate and automated analysis of brain MRI images. The proposed method therefore offers considerable potential for supporting computer-aided brain image analysis and reducing the limitations associated with conventional manual segmentation.

کلیدواژه ها:

Image Segmentation ، K-means Algorithm ، Improved Bat Optimization Algorithm