Performance Evaluation of Natural Scenes Features to create Opinion Unaware-Distortion Unaware IQA Metric

saifeldeen abdalmajeed mahmood

Abstract


There are many challenges facing image quality assessment (IQA) task. The greatest one which has been treated by this research is the difficulty of quantifying and evaluating distorted images quality blindly with no existence of the original (reference) image or partially from it. Choosing the appropriate features plays a significant role in measuring image quality. This study evaluates the efficiency of a set of features in quantifying image quality. The features have been gathered in spatial domain using the techniques of both rich edges and sharper regions of pristine natural images. The performance efficiency of these features examined through comparing them with both features gathered from reference and distorted images. These techniques employed to build two IQA metrics. Results clearly show the proposed pristine natural features competes reference features in assessing the distorted image quality. This proves the validity of these features in creating a robust metrics for evaluating distorted images. When testing the proposed metrics on LIVE database, experiment results show extracting features by means of rich edges is better than extracting it using sharper regions when assess the prediction monotonicity and applying the prediction accuracy evaluation. Besides they show the average outcome of the two techniques not only competes the popular full-reference peak signal-to-noise ratio (PSNR), the structural similarity (SSIM), and the developed NR natural image quality evaluator (NIQE) model but also outperform them.

Keywords


Key words: Key words: Key words: Key words: Key words: Key words: natural features natural featuresnatural features natural featuresnatural featuresnatural featuresnatural features natural featuresnatural features natural featuresnatural features, image ,

Full Text:

Untitled

References


SathyaVeera Reddy Dendi, ChanderDev, Narayan

Kothari, Sumohana S. Channappayya, Generating

Image Distortion Maps Using Convolutional

Autoencoders With Application to No Reference

Image Quality Assessment, Signal Processing

Letters IEEE, vol. 26, no. 1, pp. 89-93,

Jomgyoo Kim, Sanghoon Lee, Deep Image

Quality Assement By Employing FR-IQA,

International conference on Image Processing,

IEEE, 17-20, Beijing, China, Sept. 2017.

Wang, Z., A.C. Bovik, H.R. Sheikh and E.P.

Simoncelli. Image quality assessment: From

Error visibility to structural similarity. IEEE

Trans.13: 600-612, 2004.

Li, X. Blind image quality assessment.

Proceedings of the IEEE International

Conference on Image Processing, Rochester,

September 22-25, New York, USA, 2002.

Gabarda, S. and G. Cristobal. No-reference image

quality assessment through the von Mises

distribution. J. Opt. Soc. Am. A,29: 2058-2066,

Wang, Z., A.C. Bovik and B.L. Evans, Blind

measurement of blocking artifacts in

images.Proceedings of the International

Conference on Image Processing, September

-13, Vancouver, BC, 2000.

Bovik, A.C. and S. Liu. DCT-domain blind

measurement of blocking artifacts in DCT-coded

images.Proceedings of the IEEE International

Conference on Acoustics, Speech and Signal

Processing May 7-11, Salt Lake City, UT., USA,

Zhang, J., S.H. Ong and T.M. Le. Kurtosis-based

no-reference quality assessment for JPEG2000

images. Signal Process. Image Commun., 26:

-23, 2011.

Mittal, A., A.K. Moorthy and A.C.

Bovik.No-reference image quality assessment in

the spatial domain. IEEE Trans. Image

Process.21: 4695-4708, 2012.

A.K.Moorthy, and A.C. Bovik. Blind image

quality assessment: From natural scene statistics

to perceptual quality. IEEE Trans. Image

Proces.20: 3350-3364, 2011.

M.,Saad, A.C. Bovik and C. Charrier. Blind

image quality assessment: A natural scene

statistics approach in the DCT domain EEE Trans.

Image Process.21: 3339-3352, 2012.

A.,Mittal, G.S. Muralidhar, J. Ghosh and A.C.

Bovik. Blind image quality assessment without

human training using latent quality factors.IEEE

Signal Process.Lett. 19: 75–78, 2011.

A.,Mittal, R. Soundararajan and A.C. Bovik.

Making a completely blind image quality

analyzer.IEEE Signal Process.Lett.20: 209-212,

SaifeldeenAbdalmajeed and Jiao Shuhong,

No-reference image quality assessment

algorithm based on Weibull statistics of

log-derivatives of natural scenes, Electronics

Letters, Volume 50, Issue 8, , p. 595 – 596,

DOI: 10.1049/el.2013.3585, 10 April 2014.

Punit Singh and Damon M. Chandler, "F-MAD:

A feature-Based Extension of the Most Apparent

Distortion Algorithm for Image Assessment," in

Proc.of SPIE, 2013.

Metric FF Gblur WN JPEG JP2K average

SROCC

PSNR 0.7817 0.8086 0.6858 0.8478 0.8424 0.7933

SSIM 0.6535 0.6536 0.6430 0.8086 0.8014 0.7120

NIQE 0.7860 0.7602 0.8636 0.8669 0.8571 0.8268

Sharper patches 0.7882 0.7313 0.7440 0.8860 0.8549 0.8009

edges patches 0.7500 0.7628 0.78446 0.8692 0.8434 0.8020

PLCC

PSNR 0.7440 0.7572 0.7643 0.8271 0.8265 0.7778

SSIM 0.7442 0.7871 0.8193 0.7840 0.7668 0.7803

NIQE 0.8592 0.9102 0.8025 0.7520 0.8931 0.8434

Sharper patches 0.8183 0.9766 0.9009 0.8712 0.8866 0.8907

edges patches 0.9884 0.9875 0.9797 0.9074 0.8726 0.9471

Table 1: Comparison of SROCC and PLCC for the

presented algorithms versus: FR-PSNR and

FR-SSIM algorithms and NR-NIQE. Bold italic

indicates best-performing algorithm between NR

algorithms

SUST Journal of Engineering and Computer Sciences (JECS), Vol. 02, No. 3, 2019

D. L. Ruderman, The statistics of natural images.

Network: Comput. Neural Syst., 5: 517–548,

Ng KuangChern, N., Poo AunNeow and Marcelo

H. Ang Jr.”. Practical Issues in Pixel-based

Autofocusing for Machine Vision”, Proceedings

of the IEEE,3:2791 – 2796, 2001.

D., Martin, C. Fowlkes, D. Tal and J. Malik. A

database of human segmented natural images and

its application to evaluating segmentation

algorithms and measuring ecological statistics.

Proceedings of the 8th International Conference

on Computer Vision, (2001) July 7-14, ,

California University Berkeley, CA., USA.

H.R.,Sheikh, Z. Wang, L. Cormack and A.C.

Bovik, LIVE image quality assessment database

release 2, 2005. Laboratory for Image and Video

Engineering.http://live.ece.utexas.edu/research/q

uality/subjective.htm.

H.R., Sheikh, M.F. Sabir and A.C. Bovik, A

statistical evaluation of recent full reference

image quality assessment algorithms. IEEE

Trans. Image Processing,.15: 3440–3451 (2006).

Abdalmajeed, Saifeldeen&Shuhong, Jiao. Using

the Natural Scenes‟ Edges for Assessing Image

Quality Blindly and Efficiently.Mathematical

Problems in Engineering.1-9.