American Sign Language Recognition with Convolutional Neural Networks: A Gateway to Enhanced Inclusivity

In the realm of higher education, where the right to inclusive education is fundamental, it is essential to address the distinctive challenges faced by students with disabilities, particularly those who are deaf. This study explores inadequacies in accommodating the needs of deaf students within a university setting through the implementation of a Convolutional Neural Network (CNN) for sign language recognition. We conducted thorough testing of our model using three optimizers Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent (SGD), and Stochastic Gradient Descent (SGD) with Nesterov’s Momentum. Notably, the highest test accuracy of 96% was attained when employing SGD with Nesterov momentum on the Sign MNIST dataset. These findings underscore that our proposed approach outperforms other state-of-the-art methods.