American-Sign-Language-Interpeter

A program for deaf individuals that interprets sign language to written text using deep learning.

prerequisites

OpenCV2 Numpy Keras Tensorflow PYNPUT

The Model

The model has been trained on Google Colab using Convolutional Neural Network (CNN) and has reached an accuracy of 95% on test data

To test the model run "test model.py"

To obtain the the numpy classes for training run "preprocessing data.py"

Using the Model to Predict

Run "opencv_test.py" to run the program, sign different letters of the alphabet on a white background preferably. To enter a space press space on the keyboard, to delete a letter press backspace and press q to exit.