A comparison of objective evaluation algorithms for full-scale image examination

Lin Zhang and others in the paper "A comprehensiveevaluation of full REFERENCE IMAGE quality assessment algorithms". Compare some of the full-reference image quality evaluation algorithms, here to record their results.

The following table sees the library of images they use, including: Tid2008database,csiq Database,livedatabase,ivcdatabase,toyama-mictdatabase,cornell A57 Database, and wirelessimaging quality Database (WIQ).

From top to bottom, the size of the database declined in turn.

It's a total comparison. For example, the objective quality evaluation algorithm for the full-participation image is as follows:

Peaksignal to Noise Ratio (PSNR)

Peak signal-to-noise ratio.

Noise quality measure (NQM) index

References: N. Damera-venkata, T.d Kite, W.s Geisler, B.L. Evans, and A.c.bovik, "Image quality assessment based on A degradation model , "IEEE Trans. Ip,vol. 9, pp. 636-650, 2000.

Universal Quality Index (UQI)

References: Z. Wang and A.C. Bovik, "A Universal Image Quality Index," IEEE signalprocess. Lett., vol. 9, pp. 81-84, 2002.

Structural similarity (SSIM) index

References: Z. Wang, A.C Bovik, H.R Sheikh, and E.P Simoncelli, "Image qualityassessment:from error visibility to structural Similarit Y, "IEEE Trans. Ip,vol, pp. 600-612, 2004.

Multi-scalessim (Ms-ssim) index

References: Z. Wang, E.P Simoncelli, and A.C Bovik, "multi-scale structuralsimilarity for Image quality assessment," ACSSC ", pp. 1398 -1402, 2003.

Information Fidelity criterion (IFC) index

References: H.R Sheikh, A.C Bovik, and G. de Veciana, "an information fidelitycriterion for image quality assessment using Natu RAL scene Statistics, "Ieeetrans. IP, vol, pp. 2117-2128, 2005.

Visual information Fidelity (VIF) index

References: H.R Sheikh and A.C. Bovik, "Image information and visual quality," IEEE Trans. IP, vol. 430-444, pp. 2006.

Visual signal to Noise ratio (VSNR) index

References: D.M Chandler and S.S. Hemami, "vsnr:a wavelet-based visualsignal-to-noise ratio for natural images," IEEE Trans. IP, vol. 2007, pp.2284-2298.

Information Content weighted Ssim (iw-ssim) index

References: Z. Wang and Q. Li, "Information content weighting for perceptualimage quality assessment," IEEE Trans. IP, vol. 20,

pp. 1185-1198, 2011.

Riesz transforms based feature similarity (RFSIM) index

References: L. Zhang, L. Zhang, and X. Mou, "rfsim:a feature based imagequality assessment metric using Riesz transforms," Icip ", pp. 321-324, 2010.

Feature similarity (FSIM) index

References: L. Zhang, L. Zhang, X. Mou, and D. Zhang, "fsim:a feature Similarityindex for image quality assessment," IEEE Trans. IP, vol . pp. 2378-2386,2011.

The correlation coefficients between objective values and subjective values of each full-reference image quality evaluation algorithm are statistically:

Spearman rank correlation coefficient (spearman rankorder correlation coefficient,srocc), Kendall rank correlation coefficient (kendallrank-order correlation coefficient, KROCC), Pearson Linear correlation coefficient (pearsonlinear correlation COEFFICIENT,PLCC). The higher the correlation between the results of the objective algorithm and the subjective evaluation, the closer the value of the above three coefficients is to 1. Describes the more accurate the algorithm. As the table shows, the accuracy of the FSIM algorithm is relatively high, and the value of the three coefficients has reached 0.9094,0.7409,0.9050 respectively.

The table below ranks the values in the table below. There are fsim,iw-ssim in the front. Rfsim,ms-ssim.

Suddenly found: Psnr really is not allowed AH ~ ~

The following table reflects the time-consuming and time-consuming process for each of the full-test quality evaluation algorithms, which indicates that the algorithm is faster.

Overall speaking Fsim. Iw-ssim,rfsim these three kinds of new image quality evaluation algorithms are more accurate.

A comparison of objective evaluation algorithms for full-scale image examination