A MATLAB-Based Convolutional Neural Network Approach For Face Recognition System
The research on face recognition still continues after several decades since the study of this biometric trait exists. This paper discusses a method on developing a MATLAB-based Convolution Neural Network (CNN) face recognition system with Graphical User Interface (GUI) as the user input. The propos...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Ommega Online Publisher
2016
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/17280/2/A%20MATLAB-Based%20Convolutional%20Neural%20Network%20Approach%20for%20Face%20Recognition%20System.pdf http://eprints.utem.edu.my/id/eprint/17280/ http://www.ommegaonline.org/admin/journalassistance/publishimages/3259_JBPR-15-RA-009.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Teknikal Malaysia Melaka |
| Language: | English |
| Summary: | The research on face recognition still continues after several decades since the study of this biometric trait exists. This paper discusses a method on developing a MATLAB-based Convolution Neural Network (CNN) face recognition system with Graphical User Interface (GUI) as the user input. The proposed CNN has the ability to accept new subjects by training the last two layers out of four layers to reduce the neural network training time. The image preprocessing steps were implemented in MATLAB, while the CNN algorithm was implemented in C language (using GCC compiler). The main purpose of this research is to develop a complete system of face recognition. A Graphical User Interface (GUI) in MATLAB links all the steps starting from image preprocessing to face identification process. Evaluation was carried out using the images of 40 subjects from AT & T database and 10 subjects from JAFFE database producing 100% accuracy with less than 1 minute average training time when inserting 1 to 10 new subjects into the system. |
|---|