YOLOv5 Training and Configuration• Backbone: Feature Extraction• Neck: Multi-scale Features• Head: Bounding Boxes & Classes• Hyperparameters: Batch size, Learning rate, Epochs• GPU-enabled Training

Backend Implementation (Django)• User Authentication• Image Handling & Storage• REST API (Django REST Framework)

Frontend Implementation (Flet)• Dashboard (Scan Results)• Upload Panel• Settings (Sensitivity, Filters)

System Integration & Testing• Batch Image Upload• Live Camera Feed• Offline Mode with Logs• User Testing (Parents, Students)

Evaluation Metrics• Accuracy• False Positives/Negatives• System Latency• User Feedback

Limitations• No full real-time video analysis• Dataset labeling inconsistencies• Performance drop on low-power devices

ffgjh