Improve of contrast-distorted image quality assessment based on convolutional neural networks

Ismail Taha Ahmed, Soong Der Chen, Norziana Jamil, Mohamad Afendee Mohamed

Research output: Contribution to journalArticle

Abstract

Many image quality assessment algorithms (IQAs) have been developed during the past decade. However, most of them are designed for images distorted by compression, noise and blurring. There are very few IQAs designed specifically for Contrast Distorted Images (CDI), e.g. Reduced-reference Image Quality Metric for Contrast-changed images (RIQMC) and NR-IQA for Contrast-Distorted Images (NR-IQA-CDI). The existing NR-IQA-CDI relies on features designed by human or handcrafted features because considerable level of skill, domain expertise and efforts are required to design good handcrafted features. Recently, there is great advancement in machine learning with the introduction of deep learning through Convolutional Neural Networks (CNN) which enable machine to learn good features from raw image automatically without any human intervention. Therefore, it is tempting to explore the ways to transform the existing NR-IQA-CDI from using handcrafted features to machine-crafted features using deep learning, specifically Convolutional Neural Networks (CNN). The results show that NR-IQA-CDI based on non-pre-trained CNN (NR-IQA-CDI-NonPreCNN) significantly outperforms those which are based on handcrafted features. In addition to showing best performance, NR-IQA-CDI-NonPreCNN also enjoys the advantage of zero human intervention in designing feature, making it the most attractive solution for NR-IQA-CDI.

Original languageEnglish
Pages (from-to)5604-5614
Number of pages11
JournalInternational Journal of Electrical and Computer Engineering
Volume9
Issue number6
DOIs
Publication statusPublished - 01 Jan 2019

Fingerprint

Image quality
Neural networks
Learning systems
Deep learning

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Electrical and Electronic Engineering

Cite this

@article{1c5e3070da104137b1bf691cd7476043,
title = "Improve of contrast-distorted image quality assessment based on convolutional neural networks",
abstract = "Many image quality assessment algorithms (IQAs) have been developed during the past decade. However, most of them are designed for images distorted by compression, noise and blurring. There are very few IQAs designed specifically for Contrast Distorted Images (CDI), e.g. Reduced-reference Image Quality Metric for Contrast-changed images (RIQMC) and NR-IQA for Contrast-Distorted Images (NR-IQA-CDI). The existing NR-IQA-CDI relies on features designed by human or handcrafted features because considerable level of skill, domain expertise and efforts are required to design good handcrafted features. Recently, there is great advancement in machine learning with the introduction of deep learning through Convolutional Neural Networks (CNN) which enable machine to learn good features from raw image automatically without any human intervention. Therefore, it is tempting to explore the ways to transform the existing NR-IQA-CDI from using handcrafted features to machine-crafted features using deep learning, specifically Convolutional Neural Networks (CNN). The results show that NR-IQA-CDI based on non-pre-trained CNN (NR-IQA-CDI-NonPreCNN) significantly outperforms those which are based on handcrafted features. In addition to showing best performance, NR-IQA-CDI-NonPreCNN also enjoys the advantage of zero human intervention in designing feature, making it the most attractive solution for NR-IQA-CDI.",
author = "Ahmed, {Ismail Taha} and Chen, {Soong Der} and Norziana Jamil and Mohamed, {Mohamad Afendee}",
year = "2019",
month = "1",
day = "1",
doi = "10.11591/ijece.v9i6.pp5604-5614",
language = "English",
volume = "9",
pages = "5604--5614",
journal = "International Journal of Electrical and Computer Engineering",
issn = "2088-8708",
publisher = "Institute of Advanced Engineering and Science (IAES)",
number = "6",

}

Improve of contrast-distorted image quality assessment based on convolutional neural networks. / Ahmed, Ismail Taha; Chen, Soong Der; Jamil, Norziana; Mohamed, Mohamad Afendee.

In: International Journal of Electrical and Computer Engineering, Vol. 9, No. 6, 01.01.2019, p. 5604-5614.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Improve of contrast-distorted image quality assessment based on convolutional neural networks

AU - Ahmed, Ismail Taha

AU - Chen, Soong Der

AU - Jamil, Norziana

AU - Mohamed, Mohamad Afendee

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Many image quality assessment algorithms (IQAs) have been developed during the past decade. However, most of them are designed for images distorted by compression, noise and blurring. There are very few IQAs designed specifically for Contrast Distorted Images (CDI), e.g. Reduced-reference Image Quality Metric for Contrast-changed images (RIQMC) and NR-IQA for Contrast-Distorted Images (NR-IQA-CDI). The existing NR-IQA-CDI relies on features designed by human or handcrafted features because considerable level of skill, domain expertise and efforts are required to design good handcrafted features. Recently, there is great advancement in machine learning with the introduction of deep learning through Convolutional Neural Networks (CNN) which enable machine to learn good features from raw image automatically without any human intervention. Therefore, it is tempting to explore the ways to transform the existing NR-IQA-CDI from using handcrafted features to machine-crafted features using deep learning, specifically Convolutional Neural Networks (CNN). The results show that NR-IQA-CDI based on non-pre-trained CNN (NR-IQA-CDI-NonPreCNN) significantly outperforms those which are based on handcrafted features. In addition to showing best performance, NR-IQA-CDI-NonPreCNN also enjoys the advantage of zero human intervention in designing feature, making it the most attractive solution for NR-IQA-CDI.

AB - Many image quality assessment algorithms (IQAs) have been developed during the past decade. However, most of them are designed for images distorted by compression, noise and blurring. There are very few IQAs designed specifically for Contrast Distorted Images (CDI), e.g. Reduced-reference Image Quality Metric for Contrast-changed images (RIQMC) and NR-IQA for Contrast-Distorted Images (NR-IQA-CDI). The existing NR-IQA-CDI relies on features designed by human or handcrafted features because considerable level of skill, domain expertise and efforts are required to design good handcrafted features. Recently, there is great advancement in machine learning with the introduction of deep learning through Convolutional Neural Networks (CNN) which enable machine to learn good features from raw image automatically without any human intervention. Therefore, it is tempting to explore the ways to transform the existing NR-IQA-CDI from using handcrafted features to machine-crafted features using deep learning, specifically Convolutional Neural Networks (CNN). The results show that NR-IQA-CDI based on non-pre-trained CNN (NR-IQA-CDI-NonPreCNN) significantly outperforms those which are based on handcrafted features. In addition to showing best performance, NR-IQA-CDI-NonPreCNN also enjoys the advantage of zero human intervention in designing feature, making it the most attractive solution for NR-IQA-CDI.

UR - http://www.scopus.com/inward/record.url?scp=85071679155&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071679155&partnerID=8YFLogxK

U2 - 10.11591/ijece.v9i6.pp5604-5614

DO - 10.11591/ijece.v9i6.pp5604-5614

M3 - Article

VL - 9

SP - 5604

EP - 5614

JO - International Journal of Electrical and Computer Engineering

JF - International Journal of Electrical and Computer Engineering

SN - 2088-8708

IS - 6

ER -