American-Sign-Language-Classifier

Classifying Letters of the American Sign Language.

I used the VGG16 network for transfer learning.

I tried to use a scratch model but couldn't get the accuracy over 10%

Requirements

Pytorch and Numpy.

OpenCV is required for ASL-Detection.

You can find the installation instructions for Numpy here

And the one for Pytoch here

And here is the github repo for OpenCV

Where to find dataset

I got the dataset from Kaggle.

I split it into a train and test set but a validation set would also be useful.