Hand Gesture Recognition System for Recognizing of Indonesian Sign Language Number Using Attentional Convolutional Network

Abstract

Hand gestures are a form of communication widely used in everyday life. In addition, hand gestures are one of the interactions that can carry out in a non-contact manner that a computer can understand by recognizing the meaning of the movement from the image. However, hand gesture recognition cannot acknowledge directly by a computer; the computer requires artificial intelligence to recognize existing objects. For these problems, this research conducts a hand gesture recognition system for Numbers Sign of Indonesian Sign Language (SIBI) recognition using the Attentional Convolutional Network (ACN) method. Based on the results of the tests, the system can recognize number gestures indicated by a prediction accuracy value of 87.5% for non-real-time testing and accuracy value of 71.25% for real-time testing.