Image Segmentation for MRI Brain Tumor Detection Using Advance AI algorithm

• Published in: 2024 OPJU International Technology Conference (OTCON) on Smart Computing for Innovation and Advancement in Industry 4.0 pp. 1 - 8
• Main Authors: Karthikeyan, S., Lakshmanan, S.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 05.06.2024
• Subjects: Brain modeling; Data models; Deep learning; Image segmentation; Magnetic resonance imaging; MRI; Object detection; Training; YOLOv8
• DOI: 10.1109/OTCON60325.2024.10688248

The existence of a brain tumor indicates an unusual growth of cells in the brain, manifesting as either benign (non-cancerous) or malignant (cancerous). Artificial Intelligence (AI) plays a crucial role in detecting and diagnosing brain tumors, primarily utilizing medical imaging techniques such as Magnetic Resonance Imaging (MRI). This initiative aims to employ Roboflow for data annotation and model training, utilizing the YOLOv8 algorithm to precisely handle the detection of brain tumors in MRI scans through image segmentation. The dataset utilized in this project encompasses annotated MRI scans illustrating tumor regions, meticulously prepared and enhanced through the application of Roboflow. YOLOv8, renowned for its expeditious and accurate object detection capabilities, is implemented and fine-tuned utilizing the annotated dataset. Model evaluation incorporates metrics such as Mean Average Precision (MAP), Recall, and Precision. Although YOLOv8 is inherently designed for object detection, this study explores adaptations for image segmentation, treating tumors as distinct classes. The iterative refinement process encompasses fine-tuning and optimization to elevate model performance and ensure robust generalization to novel data. The focus is on the advancement of MRI-based brain tumor detection through the utilization of cutting-edge deep learning techniques. This study highlights the potential of customizing YOLOv8 for the precise task of detecting brain tumors in MRI scans, establishing a foundation for future progress in the field of medical imaging.