Brain Tumor Detection and Segmentation in MRI Images Using YOLO V۱۱

Brain tumors are among the most critical medical conditions requiring precise detection, especially at early stages. Implementing AI This study employs the YOLO V۱۱ deep learning model, a state-of-the-art fast object detection framework, to identify and segment brain tumors in MRI images. The dataset comprises ۲۰۶۴ training and ۱۰۰۰ validation images, all in ۵۱۲x۵۱۲ resolution with accompanying masks. The YOLO V۱۱ network was trained over ۱۹۳ epochs with a batch size of ۱۶. The normalized confusion matrix reveals that ۶۳% of tumor samples were correctly predicted as tumors (True Positives), while ۱۹% of the actual background samples were incorrectly classified as tumors (False Positives). Additionally, ۳۷% of tumor samples were misclassified as background (False Negatives), and ۸۱% of background samples were correctly identified as background (True Negatives). The reduced accuracy can be attributed to factors such as the limited size of the dataset, variations in tumor shapes and sizes, and the ۲D nature of the images.

کلیدواژه ها:

Brain Tumor ، Object Detection ، Segmentation ، YOLO V۱۱ ، MRI images