Journal of Systems Engineering and Electronics ›› 2019, Vol. 30 ›› Issue (2): 223-237.doi: 10.21629/JSEE.2019.02.01

• Electronics Technology •     Next Articles

No-reference image quality assessment based on AdaBoost BP neural network in wavelet domain

Junhua YAN1,2,*(), Xuehan BAI1(), Wanyi ZHANG1(), Yongqi XIAO1(), Chris CHATWIN2(), Rupert YOUNG2(), Phil BIRCH2()   

  1. 1 College of Astronautics, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
    2 School of Engineering and Informatics, University of Sussex, Brighton BN1 9QT, UK
  • Received:2017-12-19 Online:2019-04-01 Published:2019-04-26
  • Contact: Junhua YAN E-mail:yjh9758@126.com;1595720931@qq.com;daisyzwy917@126.com;couragexyq@163.com;C.R.Chatwin@sussex.ac.uk;R.C.D.Young@sussex.ac.uk;p.m.birch@sussex.ac.uk
  • About author:YAN Junhua was born in 1972. She received her B.S. degree, M.S. degree and Ph.D. degree from Nanjing University of Aeronautics and Astronautics in 1993, 2001 and 2004, respectively. She is a professor at Nanjing University of Aeronautics and Astronautics. Her research interests include image quality assessment, multi-source information fusion, target detection, tracking and recognition. E-mail:yjh9758@126.com|BAI Xuehan was born in 1993. She received her B.S. degree from Harbin Institute of Technology in 2016. She is a graduate student at Nanjing University of Aeronautics and Astronautics. Her research interest is image quality assessment. E-mail:1595720931@qq.com|ZHANG Wanyi was born in 1992. She received her B.S. degree and M.S. degree from Nanjing University of Aeronautics and Astronautics in 2014 and 2017. Her research interest is image quality assessment. E-mail:daisyzwy917@126.com|XIAO Yongqi was born in 1994. He received his B.S. degree and M.S. degree from Nanjing University of Aeronautics and Astronautics in 2015 and 2018. His research interest is image quality assessment. E-mail:couragexyq@163.com|CHATWIN Chris was born in 1950. He received his B.S. degree from University of Aston in 1973, his M.S. and Ph.D. degrees from University of Birmingham in 1977, and 1980, respectively. He is a professor in Engineering at University of Sussex. His research interests are smart cameras and algorithms, UAV surveillance systems, multi camera tracking systems, image processing for cancer diagnostics. E-mail:C.R.Chatwin@sussex.ac.uk|YOUNG Rupert was born in 1959. He received his B.S. and Ph.D. degrees from University of Glasgow in 1986, and 1994, respectively. He is a reader at University of Sussex. His current research interests are machine vision, image processing, neural networks, and quantum computing. E-mail:R.C.D.Young@sussex.ac.uk|BIRCH Phil was born in 1973. He received his B.S., M.S. degree and Ph.D. degrees from University of Durham in 1994, and 1999, respectively. He is a senior lecturer at University of Sussex. His current research interests are optical signal processing, target detection and computer vision. E-mail:p.m.birch@sussex.ac.uk
  • Supported by:
    the National Natural Science Foundation of China(61471194);the National Natural Science Foundation of China(61705104);the Science and Technology on Avionics Integration Laboratory and Aeronautical Science Foundation of China(20155552050);the Natural Science Foundation of Jiangsu Province(BK20170804);This work was supported by the National Natural Science Foundation of China (61471194; 61705104), the Science and Technology on Avionics Integration Laboratory and Aeronautical Science Foundation of China (20155552050), and the Natural Science Foundation of Jiangsu Province (BK20170804)

Abstract:

Considering the relatively poor robustness of quality scores for different types of distortion and the lack of mechanism for determining distortion types, a no-reference image quality assessment (NR-IQA) method based on the AdaBoost BP neural network in the wavelet domain (WABNN) is proposed. A 36-dimensional image feature vector is constructed by extracting natural scene statistics (NSS) features and local information entropy features of the distorted image wavelet sub-band coefficients in three scales. The ABNN classifier is obtained by learning the relationship between image features and distortion types. The ABNN scorer is obtained by learning the relationship between image features and image quality scores. A series of contrast experiments are carried out in the laboratory of image and video engineering (LIVE) database and TID2013 database. Experimental results show the high accuracy of the distinguishing distortion type, the high consistency with subjective scores and the high robustness of the method for distorted images. Experiment results also show the independence of the database and the relatively high operation efficiency of this method.

Key words: image quality assessment (IQA), AdaBoost BP neural network (ABNN), wavelet transform, natural scene statistics (NSS), local information entropy