How apodwikat/dental_anomaly_detection_graduationproject works
Project type and purpose
• This is a deep-learning-powered anomaly-detection system for dental X-ray images. It comprises:
• A research/training pipeline (Jupyter notebooks + training scripts)
• A web-based inference application (static front-end + Python REST API)
• The README confirms it’s a proof-of-concept for detecting dental anomalies with ~95.5% accuracy, using ~1,000 panoramic X-rays.
File structure
• Top-level directories:
• JupyterCodes/ – exploratory data analysis and model-dev notebooks
• Web Page/ – web application (index1.html, style1.css, main.py, restapi.py, preprocessing_utils.py, requirements.txt)
• runs/ – training logs, metrics, and model weights (detect/train24/…)
• output_folder/ – (likely) storing inference outputs or artifacts
• Key files:
• dentist_data.yaml – data schema or paths for notebooks
• main.py + restapi.py – entrypoint(s) for the inference service (Flask/FastAPI style)
• preprocessing_utils.py – image-prep utilities used by both training and inference
• requirements.txt – lists frameworks (e.g., PyTorch/YOLO, Flask/FastAPI)
README for additional insights
• Confirms deep-learning stack, potential future clinical integration, and open-collaboration ethos.
Detail without overthinking
– Capture every file’s role: e.g., show preprocessing_utils.py as a shared module between training and inference if applicable.