Object Detection Performance Analysis and Video Testing

Real-Time Inference: The YOLOv11m model was executed on high-resolution test videos to measure frames per second (FPS) and detection latency. Accuracy Validation: It was observed that the model maintained its 91% mAP50 success rate across various lighting conditions, shadows, and camera angles, while false positive detections were successfully minimized. Data Feedback Loop: Test results were utilized to identify scenarios where the model faced challenges (e.g., extreme glare) and to further refine the dataset for optimal performance.