American Sign Computer Language (ASCL) Recognizer

A real-time American Sign Language recognition system that combines advanced gesture recognition algorithms with computer vision technologies.

Overview

This project implements a gesture recognition system for American Sign Language using:

• Jackknife.py - Time series pattern recognition algorithm
• Machete.py - A segmentation technique
• MediaPipe - Hand tracking and landmark detection
• OpenCV - Computer vision and video processing

Key Features

• Real-time ASL gesture recognition
• Multi-threaded processing architecture
• 3D hand landmark tracking
• Template recording and management
• Configurable gesture matching parameters

Supported ASL Gestures

Currently recognizes the following ASL signs:

• "Forget"
• "Thank you"
• "Like"
• "No"
• "Need"

Getting Started

1. Install dependencies:

pip install -r requirements.txt

2. Launch the main recognition system:

python Scripts/main.py

4. For recording new gesture templates:

python Scripts/TemplateCrafter.py

Usage

Real-time Recognition:

• Position your hand in front of the camera
• Allow ~3 seconds for the gesture buffer to fill
• Perform ASL gestures naturally
• Recognition results appear in the console

Template Creation:

• Use TemplateCrafter.py to record new gestures
• Review recordings with frame-by-frame playback
• Save templates for recognition training

Dependencies

• OpenCV (opencv-python, opencv-contrib-python) - Video processing
• MediaPipe - Hand tracking
• NumPy - Numerical processing
• Pillow - Image processing

Development Status

Currently in active development with focus on:

• GUI 2.0 implementation
• Expanded gesture recognition set
• Performance optimization
• Template management improvements

See checklist.md for detailed development status.

Technical Details

The system uses:

• Dynamic Time Warping (DTW) for gesture matching
• MediaPipe hand landmark detection
• Multi-threaded gesture processing pipeline
• Rate-limited recognition output
• Configurable gesture confidence thresholds

References

• Jackknife Repository
• Machete Repository
• MediaPipe Documentation
• OpenCV Documentation

Known Limitations

• Template recording requires manual frame selection
• Recognition requires consistent lighting conditions
• Limited to single-hand gestures currently

Future Developments

• Two-handed gesture support
• Improved template management system
• Automated gesture segmentation
• Extended ASL vocabulary support