A lightweight AI-powered prototype for detecting potential Child Sexual Abuse Material (CSAM) in images, with perceptual hashing, encrypted logging, and a secure web interface.
- ✅ NSFW image detection using ResNet-50 (Yahoo Open NSFW model)
- 🔐 Perceptual hash generation and matching (pHash)
- 🧠 Fast classification via OpenCV's DNN module
- 🧾 AES-256 encrypted SQLite logging
- 🔍 Blurred image preview for analyst safety
- 📦 Ready for local testing and educational use
- Python 3.10+
- Flask (Web Framework)
- OpenCV (
cv2.dnn) - Pillow + ImageHash
- Cryptography (Fernet AES encryption)
csam-detector/
├── app/
│ ├── routes.py, ai_model.py, hashing.py, database.py
├── static/
├── templates/
├── model/ # Contains .caffemodel and prototxt
├── hash_db/ # Simulated hash list
├── secure_logs/ # Encrypted logs
├── uploads/ # Uploaded images (blurred view only)
└── run.py
# Install dependencies
pip install -r requirements.txt
# Run the application
python run.pyThen open your browser at:
http://127.0.0.1:5000
This tool is intended strictly for research, educational, and forensic training purposes. It does not store any real CSAM and uses only public NSFW datasets and simulated hashes.
Please use responsibly and in compliance with local laws.
Developed by Joas Antonio dos Santos
Repo: https://github.com/CyberSecurityUP/csam-detector