The project is a Machine Learning and Computer Vision based system designed to recognize Sign Language letters using a webcam. It leverages MediaPipe for hand tracking and Scikit-Learn (SVM) for classification. This project is a part of the AI Committee in IEEE AIU Student Branch.
:Tech Stack
Programming Language: Python
Machine Learning: Scikit-Learn (Support Vector Machine - SVM)
Computer Vision: OpenCV, MediaPipe
Model Persistence: Pickle
:Features
Real-time hand landmark detection using MediaPipe.
Letter prediction with probability scores.
Converts sequential predictions into words.
Dataset preprocessing and model training pipeline included.
Achieves reliable performance with normalized 3D hand landmarks.
