VSI: A Visual Saliency Induced Index for Perceptual Image Quality Assessment

Lin Zhang, Ying Shen, and Hongyu Li


This is the website for our paper "VSI: A Visual Saliency Induced Index for Perceptual Image Quality Assessment", IEEE Transactions on Image Processing, vol. 23, no. 10, pp. 4270-4281, 2014.

Perceptual image quality assessment (IQA) aims to use computational models to measure the image quality in consistent with subjective evaluations. Visual saliency (VS) has been widely studied by psychologists, neurobiologists, and computer scientists during the last decade to investigate which areas of an image will attract the most attention of the human visual system. Intuitively, VS is closely related to IQA in that suprathreshold distortions can largely affect VS maps of images. With this consideration, we propose a simple but very effective full reference IQA method by using VS. In our proposed IQA model, the role of VS is twofold. First, VS is used as a feature when computing the local quality map of the distorted image. Second, when pooling the quality score, VS is employed as a weighting function to reflect the importance of a local region. The proposed IQA index is called Visual Saliency-based Index, VSI for short. Several prominent computational VS models have been investigated in the context of IQA and the best one is chosen for VSI. Extensive experiments performed on four large scale benchmark databases demonstrate that the proposed IQA index VSI works better in terms of the prediction accuracy than all state-of-the-art IQA indices we can find while maintaining a moderate computational complexity.

Source Code

The source code can be downloaded here: VSI.m.

Evaluation Results

VSI was evaluated on 4 publicly available largescale IQA databases, including TID2013 database, TID2008 database, CSIQ database, and LIVE database. The results (in Matlab .mat format) are provided here, together with performance evaluations for future comparisons. Each result file contains a n by 2 matrix, where n denotes the number of distorted images in the database. The first column is the VSI values, and the second column is the mos/dmos values provided by the database. For example, you can use the following matlab code to calculate the SROCC and KROCC values for VSI values obtained on the TID2013 database:


matData = load('VSIOnTID2013.mat');
VSIOnTID2013 = matData.VSIOnTID2013;
VSI_TID_SROCC = corr(VSIOnTID2013(:,1), VSIOnTID2013(:,2), 'type', 'spearman');
VSI_TID_KROCC = corr(VSIOnTID2013(:,1), VSIOnTID2013(:,2), 'type', 'kendall');


The source codes to calculate the PLCC (Pearson Linear Correlation Coefficient) and RMSE (Root Mean-Squared-Error) are also provided for each database. This needs a nonlinear regression procedure which is dependant on the initialization of the parameters. We try to adjust the parameters to get a high PLCC value. For different databases, the parameter initialization may be different. The nonlinear fitting function is of the form as described in [1].

Evaluation results of VSI on seven databases are given below. Besides, for each evaluation metric, we present its weighted-average value over all the testing datasets; and the weight for each database is set as the number of distorted images in that dataset.



Nonlinear fitting code SROCC KROCC PLCC RMSE



NonlinearFittingTID2013 0.8965 0.7183 0.9000 0.5404



NonlinearFittingTID2008 0.8979 0.7123 0.8762 0.6466



NonlinearFittingCSIQ 0.9423 0.7857 0.9279 0.0979



NonlinearFittingLIVE 0.9524 0.8058 0.9482 8.6816



กก 0.9100 0.7366 0.9033 กก


[1] H.R. Sheikh, M.F. Sabir, and A.C. Bovik, "A statistical evaluation of recent full reference image quality assessment algorithms", IEEE Trans. on Image Processing, vol. 15, no. 11, pp. 3440-3451, 2006.

Created on: Dec. 02, 2013

Last update: Aug. 30, 2014