Communication and self-expression have always been difficult for people with special needs. Due to a communication gap created by the general public's inability to learn sign language, people with disabilities have little access to education and few prospects for successful work.
We examined this problem description and did our utmost to close any gaps and we took steps to improve the situation with our project's effort.
Presentation Link - check here
Step 1 : Pull the code file into any code editor supporting .ipynb files.
Step 2 : Download the dataset1 , dataset2 and dataset3 from Kaggle.
Step 3 : Install all the necessary imports mentioned in the .ipynb code files using pip(preferred installer program).
Note: If you are using google colab you can easily perform the above mentioned steps.
All the functions are already defined and made generic in the code. Also, the comments are mentioned on how to train/validate the model. You can just follow the steps mentioned in the code to easily execute the code file.
Instructions on how to run the pre-trained model on the provided sample test dataset
Step 1 : We have already provided the 3 selected model weight files for Dataset 1 in the repository.
Step 2 : Run the code with the pre-trained weights on the sample test dataset by importing them with the help of "torch.load()" function and then run the "test function" to see the results. Different weight files can also be provided by communicating with any team member.
All the source code files has already been uploaded to Github repository.
You can easily download the three datasets from Kaggle, here are the available links-