American Sign Language Alphabet Recognition and Translation

Keywords:
Convolutional Neural Network, Image processing, Transfer learning, American Sign Language

Abstract:
This paper presents a study on American Sign Language (ASL) alphabet recognition and translation. ASL is a complex language used by the Deaf community for communication. With the increasing dependency on technology in our daily lives, and with the increasing adoption of working from home, the successful inclusion of these communities in video calls and meetings is of great importance. With great advancements in Artificial Intelligence over the past years, it has now become possible to build, train and use proven to be powerful image-processing convolutional neural networks (CNN) for successful sign language recognition and translation into English text. The models have been trained and tested on a University of Exeter-derived dataset and have achieved high accuracy in this task. We have compared the results of several pre-trained models using transfer learning, as well as our own CNN. Our study shows great possibilities for improving communication between the Deaf and hearing communities.

CITATION:

IEEE format

Copy to clipboard

N. Panić, “American Sign Language Alphabet Recognition and Translation,” in Sinteza 2023 - International Scientific Conference on Information Technology, Computer Science, and Data Science, Belgrade, Singidunum University, Serbia, 2023, pp. 312-319. doi:10.15308/Sinteza-2023-312-319

APA format

Copy to clipboard

Panić, N. (2023). American Sign Language Alphabet Recognition and Translation. Paper presented at Sinteza 2023 - International Scientific Conference on Information Technology, Computer Science, and Data Science. doi:10.15308/Sinteza-2023-312-319

BibTeX format
Download

RefWorks Tagged format
Download