Development of Artificial Neural Network Architecture for Face Recognition in Real Time

Supardi, Julian and Utami, Alvi Syahrini Development of Artificial Neural Network Architecture for Face Recognition in Real Time. International Journal Machine Learning and Computing (IJMLC). ISSN 2010-3700 (In Press)

[thumbnail of Final__ICIIC_Julian-Alvi.pdf]
Preview
Text
Final__ICIIC_Julian-Alvi.pdf

Download (203kB) | Preview

Abstract

Abstracts — Face has a biological structure that is not simple. Nevertheless, research shows that some elements of the face have the geometric characteristics that can be measured. These characteristics are called face anthropometric. The existence of face anthropometric has provided significant clues for researchers to reduce the complexity of face recognition by computer. Although various methods have been developed to face recognition, but generally the system developed accepts input from a file. This condition is a one of face recognition system causes that has not been widely applied in real world. This paper presents a system that recognizes faces in real time. Artificial Neural Networks chosen as a tool for classification, to improve recognition accuracy. In this research, there are two Neural Networks used, radial basis neural network and Back-propagation neural network. The results obtained in this research shows that the accuracy of the ANN architecture that developed is still not well, which is 80%, but the Neural Network achieves convergence in 8-9 time of repetitions.

Item Type: Article
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA174.A385 Engineering design--Data processing. Manufacturing processes--Data processing. Computer integrated manufacturing systems. Manufacturing processes--Automation. CAD/CAM systems.
Divisions: 09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User: Mr. Julian Supardi
Date Deposited: 07 Oct 2019 04:31
Last Modified: 07 Oct 2019 04:31
URI: http://repository.unsri.ac.id/id/eprint/10709

Actions (login required)

View Item View Item