Sign language is a form of communication between deaf people. In Indonesia, the formal sign language is Sistem Isyarat Bahasa Indonesia or SIBI for short which is a formal sign language based on American Sign Language. However, automatic sign language recognition still faces various challenges including the complexity of hand gestures, individual variations in sign performance, and the need for real-time interpretation. These challenges make the accuracy and efficiency of sign recognition very important. To address these issues, the Vision Transformer (ViT) method can be implemented, given its advantage in capturing important features of images and its ability in processing complex computer vision tasks. ViT or Vision Transformer is an artificial neural network architecture designed for image processing or computer vision tasks. From the training results with the Vision Transformer model, the training accuracy is 100% and the validation accuracy is 92.30%.
Copyrights © 2024