I’m looking for an experienced AI and Computer Vision specialist to guide me step-by-step in building an AI-based human anomaly detection model. The focus is on both learning and implementation, covering the entire process from feature extraction to classification. The model should be capable of detecting and categorizing anomalies such as fighting, burglary, and normal behavior using datasets like UCF Crime. Key Deliverables: - Fully functional AI model for human anomaly detection. - Python scripts/notebooks for training, testing, and evaluation. - Detailed mentorship-style guidance on: - Dataset preprocessing and augmentation. - Feature extraction using deep learning backbones (e.g., ResNet, Swin Transformer, I3D). - Model architecture design and training strategies. - Performance evaluation (accuracy, F1-score, ROC, AUC). - Documentation explaining the full workflow. - (Optional) Recorded or live sessions for conceptual clarity and troubleshooting. Timeline: 4–6 weeks with structured weekly milestones and learning sessions. Faster completion is welcome if quality is maintained.