asl-svm

real-time recognition of american sign language from webcam using svm classifier and mediapipe

Dataset used

https://www.kaggle.com/datasets/grassknoted/asl-alphabet

Requirements

Install with:

pip install -r requirements.txt

Usage

Train from your own dataset with load_asl_landmarks_data() or use data.npz.

Left mouse click to register the sign, right mouse click to delete all from the sentence buffer.

Press s to begin text-to-speech on the current sentence.