Deterministic, seed-based XOR masking for RGB images. Built for fast dataset obfuscation, computer vision preprocessing, and reproducible experiments.

• Reproducible masking: the same seed always generates the same key.
• Lightweight: pure Python + NumPy + Pillow.
• Reversible: applying XOR with the same key restores the original image.
• Practical for ML workflows where visual obfuscation is needed before sharing data.

• What it is: a deterministic XOR masking tool for images.
• What it is not: a replacement for modern cryptographic standards like AES or ChaCha20.

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

python xor_single.py --input bugsbunny.jpg --seed 42 --outdir outputs

python xor_multi.py --input bugsbunny.jpg --seeds 11 22 33 --outdir outputs

• Demo animation: assets/demo.gif
• Encrypted sample: outputs/encrypted_seed42.png
• Decrypted sample: outputs/decrypted_seed42.png
• Histogram (encrypted): outputs/hist_encrypted.png

• Deep dive: docs/TECHNICAL.md
• Security policy: SECURITY.md
• Contribution guide: CONTRIBUTING.md
• Code of conduct: CODE_OF_CONDUCT.md
• GitHub operations checklist: docs/GITHUB_OPERATION_CHECKLIST.md
• Release notes draft: docs/RELEASE_NOTES_v0.1.0.md

Use this in the repo "About" field:

Seed-based XOR image encryption for deterministic dataset masking and computer vision preprocessing.

Recommended topics:

image-encryption, xor, computer-vision, dataset-anonymization, python, numpy, pillow, image-processing, reproducibility, cryptography-education

• CLI polish
• Batch dataset encryption
• Reproducible benchmark suite

See: ROADMAP.md

If this project is useful:

• Star the repository
• Watch for updates
• Open a Discussion with your use case or benchmark