American Sign Language Text Convertor

A computer-vision prototype that converts ASL hand signs into text using a CNN-based model. Includes training notebooks and a real-time webcam demo.

Overview

• Main.ipynb: training and evaluation
• create_data.ipynb: dataset preparation
• app2.py: real-time ASL to text demo
• camera.py: webcam capture utility
• model/: trained model artifacts

Setup

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Run the Demo

python app2.py

The demo will open a webcam window and display predicted letters/words.

Dataset

See Dataset link for the source of the ASL dataset.

Notes

• On macOS, you may need to install enchant dictionaries for pyenchant.
• Model artifacts are loaded from the model/ directory.