Deep Learning for Sign Language Recognition Utilizing VGG16 and ResNet50 Models
Pratham Kaushik, Eshika Jain, Kanwarpartap Singh Gill, Rahul Singh Chauhan, Hemant Singh Pokhariya · 2024
Sign Language Recognition (SLR) is necessary for establishing communication among those who are deaf or hard of hearing. This work analyses the utilization of two widely utilized deep learning models, VGG16 and ResNet50, for tasks related to SLR. By utilizing the VGG16 and ResNet50 architectures, we were able to achieve excellent accuracy rates of 99.92% and 99.95%, respectively, in accurately recognizing sign language gestures. The successful performance of such models in precisely reading movements of the hands and gestures is shown by our research, hence facilitating seamless communication for those using sign language. This research study uses advanced deep learning techniques to contribute to the consistent improvement of SLR systems, providing promising opportunities for inclusive communication and accessibility.