AI & Machine Learning

Hand-Drawn ER Diagram Digitizer

A Python/OpenCV pipeline that converts photographed hand-drawn entity-relationship diagrams into structured relationships and SQL-oriented output.

Computer VisionOpenCVER DiagramPython

Archive status: confirmed

Project Overview

A Python/OpenCV pipeline that converts photographed hand-drawn entity-relationship diagrams into structured relationships and SQL-oriented output.

Engineering Context

The project applies computer vision and document understanding to developer tools and education. The migrated archive includes 4 archived files (3.8 MB) that records the implementation and demonstration.

System Architecture

The system can be understood as an input and sensing layer, a control or processing layer, and an output, actuator, dashboard, or user-interface layer. The archived assets are retained in the organized knowledge base for future diagrams and deeper technical reconstruction.

Technologies

  • Python
  • OpenCV
  • Grayscale processing
  • Denoising
  • Contour detection
  • Shape classification

Engineering Challenges

  • Noisy camera images
  • Hand-drawn shape variation
  • Relationship interpretation

Results

  • Produced a working research prototype and retained source code

Source and Documentation

No public repository has been confirmed for this archived project.

Future Improvements

  • Add OCR, graph validation, and modern deep-learning-based symbol detection