EvilPort2/Sign-Language
A CNN model that classifies 44 American Sign Language gestures from video streams using TensorFlow and Keras.
Not currently ranked — collecting fresh signals.
star history
This repository implements a convolutional neural network trained on grayscale images to recognize gestures from American Sign Language, covering 26 alphabets, 10 numbers, and additional signs. The model was trained on 2400 images per gesture captured via OpenCV, processed at 50x50 pixel resolution. The trained Keras model is then applied to real-time video streams to predict and output the recognized sign language gesture.
Frequently asked
- What is EvilPort2/Sign-Language?
- A CNN model that classifies 44 American Sign Language gestures from video streams using TensorFlow and Keras.
- Is Sign-Language open source?
- Yes — EvilPort2/Sign-Language is an open-source project tracked on heatdrop.
- How popular is Sign-Language?
- EvilPort2/Sign-Language has 824 stars on GitHub.
- Where can I find Sign-Language?
- EvilPort2/Sign-Language is on GitHub at https://github.com/EvilPort2/Sign-Language.