Brain Tumor Detection Using YOLOv11 on Edge Computing for Decision Support

Early diagnosis of brain tumors is essential for timely treatment and improved patient outcomes. With the evolution of deep learning, object detection models have emerged as valuable tools to assist clinical decision-making. This study investigates the deployment of YOLOv11 models (n, s, m, l, x) on a Raspberry Pi 5 with 8GB of RAM for the detection of brain tumors in magnetic resonance imaging. The goal is to evaluate the trade-offs between detection accuracy and computational efficiency in edge computing scenarios. Results show that smaller models achieve inference times under 110 milliseconds with precision above 93%, while larger variants exceed 97% precision with increased latency. The experimental evidence confirms the feasibility of real-time medical image analysis using embedded systems and highlight the robustness of YOLOv11 models even under hardware constraints.