Sudoku Vision

Scan sudoku puzzles with your camera and solve them instantly using computer vision and machine learning.

Project Structure

sudoku-vision/
├── solver/     # C sudoku solver (compiles to WASM)
├── ml/         # Python ML training pipeline
├── web/        # TypeScript frontend app
├── data/       # Training images and test fixtures
└── ROADMAP.md  # Development plan and architecture

Status

Early development. See ROADMAP.md for the full plan.

Components

• Solver: Backtracking algorithm in C, targeting WebAssembly
• ML Pipeline: CNN for digit recognition (PyTorch)
• Grid Detection: OpenCV-based cell extraction
• Web App: Camera capture, inference, solution display

License

MIT