AI-Based ASL Translation: A Study of Deep Learning and 3D Modeling

Deafness is a significant condition that can have a profound impact on the daily lives of those affected. One of the major challenges that deaf individuals face is communicating with hearing people, which can lead to feelings of isolation and even depression. In this paper, we propose a solution to bridge the communication gap between deaf and hearing individuals by utilizing a deep learning-based computer vision model to translate American Sign Language (ASL) into text and a 3D modeling to translate text, speech, or video into ASL. To achieve this goal, we integrated the Inception3D model within a convolutional neural network to interpret videos of ASL for the deaf through the application of deep learning techniques. The proposed neural network was trained using a dataset of 21, 083 videos of ASL, comprising 2, 000 commonly used words in the language. The results on the testing data showed an accuracy of \(42.69\%\) . This approach has the potential to greatly improve communication between deaf and hearing individuals and reduce feelings of isolation and depression.