Brain magnetic resonance image (MRI) segmentation using multimodal optimization

• DOI: 10.1007/s11042-024-19725-4
• Publication Date: 2024-07-02T10:03:08Z
• Authors: Taymaz Akan; Amin Golzari Oskouei; Sait Alp; Mohammad Alfrad N...
• Keywords: Real-time MRI

Abstract One of the highly focused areas in medical science community is segmenting tumors from brain magnetic resonance imaging (MRI). The diagnosis malignant at an early stage necessary to provide treatment for patients. patient’s prognosis will improve if it detected early. Medical experts use a manual method segmentation when making tumors. This study proposes new approach simplify and automate this process. In recent research, multi-level has been widely used image analysis, effectiveness precision are directly tied number segments used. However, choosing appropriate often left up user challenging many algorithms. proposed modified version 3D Histogram-based method, which can automatically determine segments. general algorithm contains three main steps: first step Gaussian filter smooth RGB histogram image. eliminates unreliable non-dominating peaks that too close together. Next, multimodal particle swarm optimization identifies histogram’s peaks. end, pixels placed cluster best fits their characteristics based on non-Euclidean distance. applied Cancer Imaging Archive (TCIA) MRI Images Tumor detection dataset. results compared with those clustering methods: FCM, FCM_FWCW, FCM_FW. comparative analysis algorithms across various slices. Our consistently demonstrates superior performance. It achieves top mean rank all metrics, indicating its robustness clustering. effective experiments, proving capacity find proper clusters.