Hand Sign Recognition

A computer vision system for real-time hand sign classification.

Overview

Developed a real-time Hand Sign Recognition system capable of identifying and classifying hand gestures.

Methodology

  • Computer Vision Techniques: Utilized OpenCV for image pre-processing, including:
    • Skin color segmentation.
    • Convex Hull generation to outline the hand shape.
    • Convexity Defects calculation to count fingers and analyze gestures.
  • Classification: Experimented with both heuristic-based approaches (finger counting) and Neural Networks for robust classification of complex signs.

Tech Stack

  • OpenCV
  • Python
  • Machine Learning