Back to Projects
AI / Deep Learning

Dynamic Sign Language Recognition

Developed a system for recognizing dynamic (motion-based) sign language gestures in real-time using a combination of MediaPipe for pose estimation, LSTM networks for temporal modeling, and Dynamic Time Warping for sequence matching.

MediaPipeLSTMDTWPythonTensorFlowOpenCV
GitHub Repository
IEEE GITCON 2026
🎯

The Problem

Static sign recognition systems cannot handle dynamic gestures that involve motion and temporal patterns.

💡

The Solution

Combined MediaPipe skeletal tracking with LSTM temporal modeling and DTW sequence matching for robust dynamic gesture recognition.

// OUTCOMES

Key Results

Real-time dynamic gesture recognition
Robust temporal pattern matching
Published at IEEE GITCON 2026