This repository contains software and data for performing sign language translation training and testing on a mnist sign language dataset. Additionally, there is support for a live camera feed sign language translation.
You will need to install sklearn, numpy, pytorch, and onnx
The data used is taken from a kaggle dataset called Sign Language MNIST which is an MNIST adaptation of a sign language dataset.
Running the system can be split up into 3 steps:
- To train the network run train.py where you can specify any of the models to be used that can be found in networks.py.
- Running evaluation.py gives you the results of testing the network on the test set and also create a production version of the model with onnx.
- For a live feed where it translate your hand signal in real-time you can run the live_translation.py file after doing the previous two steps.
This repository is created by Alec Nonnemaker and Sara Boby. Feel free to email us if you have any questions at [email protected] or [email protected]