Deep Learning-Based Classification of Magnetic Resonance Brain Images using YOLOv5 and Lion Swarm Optimization | ||||
| Menoufia Journal of Electronic Engineering Research | ||||
| Volume 34, Issue 2, July 2025, Page 12-19 PDF (572.41 K) | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/mjeer.2025.325982.1096 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
Zeinab F. Elsharkawy   1; Ahmed S. Elkorany2; Said M. Elhalafawy2
| ||||
| 1Engineering Department, Nuclear Research Center, Egyptian Atomic Energy Authority(EAEA) | ||||
| 2Dept. of Electronics and Electrical Comm. Eng., Faculty of Electronic Engineering, Menoufia University, Menouf, 32952, Egypt. | ||||
| Abstract | ||||
| Accurate classification of the magnetic resonance (MR) brain images is crucial for diagnosis and therapy planning. This research presents an efficient approach for MR brain image diagnoses by combining YOLOv5, the revolutionary object detection technique, with Lion Swarm Optimization (LSO). LSO, chosen for its high optimization precision, rapid convergence, and robust stability, improves YOLOv5 by efficiently fine-tuning hyperparameters and optimizing decision thresholds, resulting in higher classification accuracy (CA) and faster convergence during training. The goal is to accurately classify MR brain images into low- and high- grade gliomas (LGG & HGG) categories. YOLOv5 is employed in the suggested method to automatically identify regions of interest (ROIs) that are suggestive of HGG or LGG by detecting and localizing relevant features within MR brain images. The approach considerably increases efficiency and lowers human labor by eliminating the requirement for manual ROI selection by including YOLOv5 in the framework. To further improve the CA, the YOLOv5 model is optimized using LSO. The BRATS dataset is used to assess the suggested approach. The usefulness of the suggested approach is demonstrated by experimental findings, which outperform the other deep learning architectures and achieve an impressive Recall, Specificity, F1 score, and CA of 99%. The high accuracy highlights the potential of the proposed method as a reliable tool in clinical settings for accurate tumor classification and treatment decision-making. | ||||
| Keywords | ||||
| Deep learning; Brain Image classification; YOLOv5; Lion Swarm Optimization | ||||
|
Statistics Article View: 116 PDF Download: 93 |
||||
