Text this: Leveraging 3D skeleton video extraction and deep learning for real-time sign language recognition model