Open AccessComplex number‐based image quality assessment using singular value decomposition
Author(s) -
Yong Wang,
Yuqing Wang,
Xiaohui Zhao
Publication year - 2016
Publication title -
iet image processing
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.401
H-Index - 45
eISSN - 1751-9667
pISSN - 1751-9659
DOI - 10.1049/iet-ipr.2014.0937
Subject(s) - image quality , metric (unit) , distortion (music) , singular value decomposition , computer science , artificial intelligence , image (mathematics) , scalar (mathematics) , similarity (geometry) , mathematics , pattern recognition (psychology) , data mining , algorithm , geometry , amplifier , computer network , operations management , bandwidth (computing) , economics
In this study, the combining strategies are considered to be effective tools to improve the performance of the image quality assessment (IQA) metrics. A new metric is proposed to evaluate the quality of test images of combined degradation and individual degradation. The complex numbers are used to describe the image structure in the proposed method. On the basis of that, the properties of the classical IQA method based on singular value decomposition are analysed. The difference between energy visual map and structure visual map is shown. The complex‐number‐based approach is different from the classical scalar‐based techniques, which are insufficient to describe image structure. The proposed C_SVDQ metric can be considered as a vectorial expansion of structure similarity. In the experiments, an extensive comparison between the proposed C_SVDQ and other IQA metrics on image quality database was performed. Both the overall tests and the individual distortion tests show the superiority of this new approach in IQA.