Can convolution neural network (CNN) triumph in ear recognition of uniform illumination invariant?

Nursuriati Jamil, Ali Abd Almisreb, Syed Mohd Zahid Syed Zainal Ariffin, Norashidah Md Din, Raseeda Hamzah

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Current deep convolution neural network (CNN) has shown to achieve superior performance on a number of computer vision tasks such as image recognition, classification and object detection. The deep network was also tested for view-invariance, robustness and illumination invariance. However, the CNN architecture has thus far only been tested on non-uniform illumination invariant. Can CNN perform equally well for very underexposed or overexposed images or known as uniform illumination invariant? This is the gap that we are addressing in this paper. In our work, we collected ear images under different uniform illumination conditions with lumens or lux values ranging from 2 lux to 10,700 lux. A total of 1,100 left and right ear images from 55 subjects are captured under natural illumination conditions. As CNN requires considerably large amount of data, the ear images are further rotated at every 5o angles to generate 25,300 images. For each subject, 50 images are used as validation/testing dataset, while the remaining images are used as training datasets. Our proposed CNN model is then trained from scratch and validation and testing results showed recognition accuracy of 97%. The results showed that 100% accuracy is achieved for images with lumens ranging above 30 but having problem with lumens less than 10 lux.

Original languageEnglish
Pages (from-to)558-566
Number of pages9
JournalIndonesian Journal of Electrical Engineering and Computer Science
Volume11
Issue number2
DOIs
Publication statusPublished - 01 Aug 2018

Fingerprint

Convolution
Illumination
Lighting
Neural Networks
Neural networks
Invariant
Invariance
Image recognition
Testing
Network architecture
Robustness (control systems)
Computer vision
Image Recognition
Object Detection
Network Architecture
Neural Network Model
Computer Vision
Robustness
Angle

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Information Systems
  • Hardware and Architecture
  • Computer Networks and Communications
  • Control and Optimization
  • Electrical and Electronic Engineering

Cite this

Jamil, Nursuriati ; Almisreb, Ali Abd ; Ariffin, Syed Mohd Zahid Syed Zainal ; Md Din, Norashidah ; Hamzah, Raseeda. / Can convolution neural network (CNN) triumph in ear recognition of uniform illumination invariant?. In: Indonesian Journal of Electrical Engineering and Computer Science. 2018 ; Vol. 11, No. 2. pp. 558-566.
@article{6845ead6f32c46bdbb7c9dd54bfe3c24,
title = "Can convolution neural network (CNN) triumph in ear recognition of uniform illumination invariant?",
abstract = "Current deep convolution neural network (CNN) has shown to achieve superior performance on a number of computer vision tasks such as image recognition, classification and object detection. The deep network was also tested for view-invariance, robustness and illumination invariance. However, the CNN architecture has thus far only been tested on non-uniform illumination invariant. Can CNN perform equally well for very underexposed or overexposed images or known as uniform illumination invariant? This is the gap that we are addressing in this paper. In our work, we collected ear images under different uniform illumination conditions with lumens or lux values ranging from 2 lux to 10,700 lux. A total of 1,100 left and right ear images from 55 subjects are captured under natural illumination conditions. As CNN requires considerably large amount of data, the ear images are further rotated at every 5o angles to generate 25,300 images. For each subject, 50 images are used as validation/testing dataset, while the remaining images are used as training datasets. Our proposed CNN model is then trained from scratch and validation and testing results showed recognition accuracy of 97{\%}. The results showed that 100{\%} accuracy is achieved for images with lumens ranging above 30 but having problem with lumens less than 10 lux.",
author = "Nursuriati Jamil and Almisreb, {Ali Abd} and Ariffin, {Syed Mohd Zahid Syed Zainal} and {Md Din}, Norashidah and Raseeda Hamzah",
year = "2018",
month = "8",
day = "1",
doi = "10.11591/ijeecs.v11.i2.pp558-566",
language = "English",
volume = "11",
pages = "558--566",
journal = "Indonesian Journal of Electrical Engineering and Computer Science",
issn = "2502-4752",
publisher = "Institute of Advanced Engineering and Science (IAES)",
number = "2",

}

Can convolution neural network (CNN) triumph in ear recognition of uniform illumination invariant? / Jamil, Nursuriati; Almisreb, Ali Abd; Ariffin, Syed Mohd Zahid Syed Zainal; Md Din, Norashidah; Hamzah, Raseeda.

In: Indonesian Journal of Electrical Engineering and Computer Science, Vol. 11, No. 2, 01.08.2018, p. 558-566.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Can convolution neural network (CNN) triumph in ear recognition of uniform illumination invariant?

AU - Jamil, Nursuriati

AU - Almisreb, Ali Abd

AU - Ariffin, Syed Mohd Zahid Syed Zainal

AU - Md Din, Norashidah

AU - Hamzah, Raseeda

PY - 2018/8/1

Y1 - 2018/8/1

N2 - Current deep convolution neural network (CNN) has shown to achieve superior performance on a number of computer vision tasks such as image recognition, classification and object detection. The deep network was also tested for view-invariance, robustness and illumination invariance. However, the CNN architecture has thus far only been tested on non-uniform illumination invariant. Can CNN perform equally well for very underexposed or overexposed images or known as uniform illumination invariant? This is the gap that we are addressing in this paper. In our work, we collected ear images under different uniform illumination conditions with lumens or lux values ranging from 2 lux to 10,700 lux. A total of 1,100 left and right ear images from 55 subjects are captured under natural illumination conditions. As CNN requires considerably large amount of data, the ear images are further rotated at every 5o angles to generate 25,300 images. For each subject, 50 images are used as validation/testing dataset, while the remaining images are used as training datasets. Our proposed CNN model is then trained from scratch and validation and testing results showed recognition accuracy of 97%. The results showed that 100% accuracy is achieved for images with lumens ranging above 30 but having problem with lumens less than 10 lux.

AB - Current deep convolution neural network (CNN) has shown to achieve superior performance on a number of computer vision tasks such as image recognition, classification and object detection. The deep network was also tested for view-invariance, robustness and illumination invariance. However, the CNN architecture has thus far only been tested on non-uniform illumination invariant. Can CNN perform equally well for very underexposed or overexposed images or known as uniform illumination invariant? This is the gap that we are addressing in this paper. In our work, we collected ear images under different uniform illumination conditions with lumens or lux values ranging from 2 lux to 10,700 lux. A total of 1,100 left and right ear images from 55 subjects are captured under natural illumination conditions. As CNN requires considerably large amount of data, the ear images are further rotated at every 5o angles to generate 25,300 images. For each subject, 50 images are used as validation/testing dataset, while the remaining images are used as training datasets. Our proposed CNN model is then trained from scratch and validation and testing results showed recognition accuracy of 97%. The results showed that 100% accuracy is achieved for images with lumens ranging above 30 but having problem with lumens less than 10 lux.

UR - http://www.scopus.com/inward/record.url?scp=85048180302&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048180302&partnerID=8YFLogxK

U2 - 10.11591/ijeecs.v11.i2.pp558-566

DO - 10.11591/ijeecs.v11.i2.pp558-566

M3 - Article

AN - SCOPUS:85048180302

VL - 11

SP - 558

EP - 566

JO - Indonesian Journal of Electrical Engineering and Computer Science

JF - Indonesian Journal of Electrical Engineering and Computer Science

SN - 2502-4752

IS - 2

ER -