Generation of Synthetic Datasets for Sign Language

The project at hand is inserted in the Deep Learning study case, namely in the deep learning model training using image recognition. In order to test the learning capacities of such models, this project will focus on the image recognition of different letters in American Sign Language(ASL) and the generation of synthetic datasets. The main goal is to generate synthetic datasets using a 3D hand model and see if they can be used for sign language recognition on par with real datasets.

Authors

• Luís Pereira
• Tiago Sousa
• Ariana Lousada

Real Dataset Used:

• https://www.kaggle.com/datasets/ayuraj/asl-dataset

3D Hand Model:

• https://free3d.com/3d-model/rigged-hands-28855.html