Two Low-Level Feature Distributions Based No Reference Image Quality Assessment
No reference image quality assessment (NR IQA) aims to develop quantitative measures to automatically and accurately estimate perceptual image quality without any prior information about the reference image. In this paper, we introduce two low-level feature distributions (TLLFD) based method for NR...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-05-01
|
Series: | Applied Sciences |
Subjects: | |
Online Access: | https://www.mdpi.com/2076-3417/12/10/4975 |
_version_ | 1797501818944618496 |
---|---|
author | Hao Fu Guojun Liu Xiaoqin Yang Lili Wei Lixia Yang |
author_facet | Hao Fu Guojun Liu Xiaoqin Yang Lili Wei Lixia Yang |
author_sort | Hao Fu |
collection | DOAJ |
description | No reference image quality assessment (NR IQA) aims to develop quantitative measures to automatically and accurately estimate perceptual image quality without any prior information about the reference image. In this paper, we introduce two low-level feature distributions (TLLFD) based method for NR IQA. Different from the deep learning method, the proposed method characterizes image quality with the distributions of low-level features, thus it has few parameters, simple model, high efficiency, and strong robustness. First, the texture change of distorted image is extracted by the weighted histogram of generalized local binary pattern. Second, the Weibull distribution of gradient is extracted to represent the structural change of the distorted image. Furthermore, support vector regression is adopted to model the complex nonlinear relationship between feature space and quality measure. Finally, numerical tests are performed on LIVE, CISQ, MICT, and TID2008 standard databases for five different distortion categories JPEG2000 (JP2K), JPEG, White Noise (WN), Gaussian Blur (GB), and Fast Fading (FF). The experimental results indicate that TLLFD method achieves superior performance and strong generalization for image quality prediction as compared to state-of-the-art full-reference, no reference, and even deep learning IQA methods. |
first_indexed | 2024-03-10T03:24:05Z |
format | Article |
id | doaj.art-118bddcd5a5c44f59cc2042ba01df3f9 |
institution | Directory Open Access Journal |
issn | 2076-3417 |
language | English |
last_indexed | 2024-03-10T03:24:05Z |
publishDate | 2022-05-01 |
publisher | MDPI AG |
record_format | Article |
series | Applied Sciences |
spelling | doaj.art-118bddcd5a5c44f59cc2042ba01df3f92023-11-23T09:55:53ZengMDPI AGApplied Sciences2076-34172022-05-011210497510.3390/app12104975Two Low-Level Feature Distributions Based No Reference Image Quality AssessmentHao Fu0Guojun Liu1Xiaoqin Yang2Lili Wei3Lixia Yang4School of Mathematics and Statistics, Ningxia University, Yinchuan 750021, ChinaSchool of Mathematics and Statistics, Ningxia University, Yinchuan 750021, ChinaSchool of Mathematics and Statistics, Ningxia University, Yinchuan 750021, ChinaSchool of Mathematics and Statistics, Ningxia University, Yinchuan 750021, ChinaSchool of Mathematics and Statistics, Ningxia University, Yinchuan 750021, ChinaNo reference image quality assessment (NR IQA) aims to develop quantitative measures to automatically and accurately estimate perceptual image quality without any prior information about the reference image. In this paper, we introduce two low-level feature distributions (TLLFD) based method for NR IQA. Different from the deep learning method, the proposed method characterizes image quality with the distributions of low-level features, thus it has few parameters, simple model, high efficiency, and strong robustness. First, the texture change of distorted image is extracted by the weighted histogram of generalized local binary pattern. Second, the Weibull distribution of gradient is extracted to represent the structural change of the distorted image. Furthermore, support vector regression is adopted to model the complex nonlinear relationship between feature space and quality measure. Finally, numerical tests are performed on LIVE, CISQ, MICT, and TID2008 standard databases for five different distortion categories JPEG2000 (JP2K), JPEG, White Noise (WN), Gaussian Blur (GB), and Fast Fading (FF). The experimental results indicate that TLLFD method achieves superior performance and strong generalization for image quality prediction as compared to state-of-the-art full-reference, no reference, and even deep learning IQA methods.https://www.mdpi.com/2076-3417/12/10/4975no reference image quality assessmentlow-level featuregeneralized local binary patterngradientdeep learning |
spellingShingle | Hao Fu Guojun Liu Xiaoqin Yang Lili Wei Lixia Yang Two Low-Level Feature Distributions Based No Reference Image Quality Assessment Applied Sciences no reference image quality assessment low-level feature generalized local binary pattern gradient deep learning |
title | Two Low-Level Feature Distributions Based No Reference Image Quality Assessment |
title_full | Two Low-Level Feature Distributions Based No Reference Image Quality Assessment |
title_fullStr | Two Low-Level Feature Distributions Based No Reference Image Quality Assessment |
title_full_unstemmed | Two Low-Level Feature Distributions Based No Reference Image Quality Assessment |
title_short | Two Low-Level Feature Distributions Based No Reference Image Quality Assessment |
title_sort | two low level feature distributions based no reference image quality assessment |
topic | no reference image quality assessment low-level feature generalized local binary pattern gradient deep learning |
url | https://www.mdpi.com/2076-3417/12/10/4975 |
work_keys_str_mv | AT haofu twolowlevelfeaturedistributionsbasednoreferenceimagequalityassessment AT guojunliu twolowlevelfeaturedistributionsbasednoreferenceimagequalityassessment AT xiaoqinyang twolowlevelfeaturedistributionsbasednoreferenceimagequalityassessment AT liliwei twolowlevelfeaturedistributionsbasednoreferenceimagequalityassessment AT lixiayang twolowlevelfeaturedistributionsbasednoreferenceimagequalityassessment |