Volume 18, Issue 2 (10-2021)
JSDP 2021, 18(2): 115-134 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Daryanavard M, Shahbahrami A. Non-distortion-specific no-reference Image Quality Assessment using Statistical Features. JSDP 2021; 18 (2) :115-134
URL: http://jsdp.rcisp.ac.ir/article-1-1050-en.html
URL: http://jsdp.rcisp.ac.ir/article-1-1050-en.html
Department of Computer Engineering, Engineering Faculty, University of Guilan
Abstract: (2132 Views)
Objective Image Quality Assessment (IQA) algorithms are divided into three categories according to the availability of the reference image and the amount of information available in the assessment, namely, algorithms with Full Reference (FR), Reduced Reference (RR), and No Reference (NR). If the original high-quality image exists in the evaluation algorithm at the same time to compare with the test image, and both images are identical in content, the evaluation is called FR assessment. If only a few features extracted from the high-quality image are used to compare with the test image, it is called RR assessment. In NR assessment algorithms, there is no feature or reference image to compare with the test image. Algorithms in NR are divided into two subcategories, Distortion-Specific (DS) and Non-Distortion-Specific (NDS). In first one, algorithms predict the quality of an image by knowing the type of distortion which is effective when distortion information or type is available. However, information about the type of distortion is not available in most applications which limits the use of these algorithms. The NDS algorithms can be applied to different types of distortion and are designed to be all-purpose. The NDS algorithms are divided into two subcategories namely, Opinion Aware (OA) and Opinion Unaware (OU). In the OA model, the images are evaluated and scored by the human factor, and each image with its corresponding human score is mapped by the learning system, while the OU model does not have the score of a human observer and is evaluated completely blind.
The proposed algorithm is this paper is for NR model and NDS and opinion unaware. Image quality is well correlated with features of local structure, contrast, and color. By modeling, these features with distributions such as Gaussian or Gaussian families can be used to detect image degradation. The proposed method consists of two stages of training and testing. Five NSS features that are actually extracted from the MSCN coefficient. The q-Gaussian distribution model is used for image distribution. The q-Gaussian distribution is one of the options that create flexible decision boundaries with different Gaussian shapes that are more generalizable in anomalies than other distributions. The learning phase is performed only once to extract the features of images and to be considered as a model in the system to compare with the images that are to be entered as test images.
To evaluate the performance, the proposed method is compared with IL-NIQE which is similar to the proposed method in terms of behavioral mechanism and it uses natural scene statistics. Performance metrics such as PLCC, KROCC, RMSE, SROCC and some datasets, LIVE, CSIQ, and TID2013 have been used for evaluation. The proposed method performs better than the compared technique. The proposed method can show better performance due to the adjustable parameter of q.
The proposed algorithm is this paper is for NR model and NDS and opinion unaware. Image quality is well correlated with features of local structure, contrast, and color. By modeling, these features with distributions such as Gaussian or Gaussian families can be used to detect image degradation. The proposed method consists of two stages of training and testing. Five NSS features that are actually extracted from the MSCN coefficient. The q-Gaussian distribution model is used for image distribution. The q-Gaussian distribution is one of the options that create flexible decision boundaries with different Gaussian shapes that are more generalizable in anomalies than other distributions. The learning phase is performed only once to extract the features of images and to be considered as a model in the system to compare with the images that are to be entered as test images.
To evaluate the performance, the proposed method is compared with IL-NIQE which is similar to the proposed method in terms of behavioral mechanism and it uses natural scene statistics. Performance metrics such as PLCC, KROCC, RMSE, SROCC and some datasets, LIVE, CSIQ, and TID2013 have been used for evaluation. The proposed method performs better than the compared technique. The proposed method can show better performance due to the adjustable parameter of q.
Keywords: Image quality assessment, No-reference assessment, Non-distortion-specific assessment, q-Gaussian distribution
Type of Study: Research |
Subject:
Paper
Received: 2019/07/21 | Accepted: 2020/04/29 | Published: 2021/10/8 | ePublished: 2021/10/8
Received: 2019/07/21 | Accepted: 2020/04/29 | Published: 2021/10/8 | ePublished: 2021/10/8
References
1. [1] A. Attar, A. Shahbahrami, R. Moradi rad, "Image quality assessment using edge based features", Multimedia tools and applications, Vol. 75, pp. 7407-7422, 2015. [DOI:10.1007/s11042-015-2663-9]
2. [2] M. Saad, A. C. Bovik, and C. Charrier, "Blind image quality assessment: A natural scene statistics approach in the DCT domain", IEEE Trans.Image Process, Vol. 21, pp. 3339-3352, 2012. [DOI:10.1109/TIP.2012.2191563] [PMID]
3. [3] P. Ye and D. Doermann, "No-reference image quality assessment using visual codebooks", in Proc. IEEE Int. Conf. Image Process, Vol. 21, pp. 3129-3138, 2011. [DOI:10.1109/TIP.2012.2190086] [PMID]
4. [4] J. Shen, Q. Li, and G. Erlebacher, "Hybrid no-reference natural image quality assessment of noisy, blurry, JPEG2000, and JPEG images", IEEE Trans. Image Process, Vol. 20, pp. 2089-2098, 2011. [DOI:10.1109/TIP.2011.2108661] [PMID]
5. [5] H. R. Sheikh and A. C. Bovik, "Image information and visual quality", IEEE Trans. Image Process, Vol. 15, pp. 430-444, 2006. [DOI:10.1109/TIP.2005.859378] [PMID]
6. [6] A. K. Moorthy and A. C. Bovik, "A two-step framework for constructing blind image quality indices", IEEE Signal Process. Lett, Vol. 17, pp. 513-516, 2010. [DOI:10.1109/LSP.2010.2043888]
7. [7] H. R. Sheikh, A. C. Bovik, and L. K. Cormack, "No-reference quality assessment using natural scene statistics: JPEG2000", IEEE Trans. Image Process, Vol. 14, pp. 1918-1927, 2005. [DOI:10.1109/TIP.2005.854492] [PMID]
8. [8] M. Zandifar, J. Tahmoresnezhad, "Sample-oriented Domain Adaptation for Image Classification", Joural Signal and Data Processsing, Vol. 16, Vol. 3, pp.129-148, 2019. [DOI:10.29252/jsdp.16.3.148]
9. [9] X. Shaoping, J. Shunliang, M. Weidong, "No-reference/blind image quality assessment: a survey", Journal IETE Technical Review, Vol. 34, pp.223-245, 2016. [DOI:10.1080/02564602.2016.1151385]
10. [10] M. A. Saad, A. C. Bovik, C. Charrier, "A dct statistics-based blind image quality index", IEEE Signal Processing Letters, Vol. 17, pp.583-586, 2010. [DOI:10.1109/LSP.2010.2045550]
11. [11] M. A. Saad, A. C. Bovik, C. Charrier, "Blind image quality assessment: A natural scene statistics approach in the DCT domain", IEEE Trans. Image Processing, Vol. 21, pp.3339- 3352, 2012. [DOI:10.1109/TIP.2012.2191563] [PMID]
12. [12] R. C. Gonzalez and R. E. Woods, Digital Image Processing, 2nd. Upper Saddle River, NJ: Prentice-Hall, 2002.
13. [13] Z. Wang, A. C. Bovik, H. R. Sheikh and E. P. Simoncelli, "Image Quality Assessment: From Error Visibility to Structural Similarity", IEEE Trans. Image Process., Vol. 13, No. 4, pp. 600-612, 2004. [DOI:10.1109/TIP.2003.819861] [PMID]
14. [14] A. Mittal, A. K. Moorthy, and A. C. Bovik, "No-reference image quality assessment in the spatial domain", IEEE Trans. Image Process., Vol. 21, pp. 4695 708, 2012. [DOI:10.1109/TIP.2012.2214050] [PMID]
15. [15] T. Huixuan, J. Neel, and K. Ashish. "Learning a blind measure of perceptual image quality", IEEE Conf. on Computer Vision and Pattern Recognition, pp. 305- 312, 2011.
16. [16] P. Ye and D. Doermann "No-reference image quality assessment using visual codebooks", IEEE Trans. Image Processing, Vol. 21, pp.3129-3138, 2012. [DOI:10.1109/TIP.2012.2190086] [PMID]
17. [17] Y. Zhang and D. M. Chandler "An algorithm for no-reference image quality assessment based on log-derivative statistics of natural scenes", in IS&T/SPIE Electronic Imaging, Int. Society for Optics and Photonics, pp. 86530J-0 - 86530J-10. 2013.
18. [18] M. Zhang, X. Jin, Z. Xiangrong and F. Hiroshi, "No reference image quality assessment based on local binary pattern statistics", In Visual Communications and Image Processing, pp. 1-6, 2013. [DOI:10.1109/VCIP.2013.6706418]
19. [19] Q. Li, W. Lin, J. Xu, and Y. Fang, "Blind image quality assessment using statistical structural and luminance features", IEEE Trans. Multimedia, Vol. PP, pp. 1-1, 2016. [DOI:10.1109/TMM.2016.2601028]
20. [20] I. Nenakhov, V. Khryashchev, A. Priorov, "No-reference image quality assessment based on local binary patterns", IEEE East-West Design & Test Symposium, 2016. [DOI:10.1109/EWDTS.2016.7807685]
21. [21] W. Xue, X. Mou, L. Zhang, A. C. Bovik and X. Feng "Blind image quality assessment using joint statistics of gradient magnitude and laplacian features", IEEE Trans. Image Process, Vol.23, pp.4850-4862, 2014. [DOI:10.1109/TIP.2014.2355716] [PMID]
22. [22] H. Hadizadeh, I. V. Bajić, "Color Gaussian Jet Features for No-Reference Quality Assessment of Multiply-Distorted Images", IEEE Signal Processing Letters, vol. 23, pp. 1717 - 1721, 2016. [DOI:10.1109/LSP.2016.2617743]
23. [23] C. Sun, H. Li, W. Li, "No-reference image quality assessment based on global and local content perception", In Visual Communications and Image Processing, 2016. [DOI:10.1109/VCIP.2016.7805544]
24. [24] D. Lee, K. N. Plataniotis, "Toward a No-Reference Image Quality Assessment Using Statistics of Perceptual Color Descriptors", IEEE Trans. Image Processing, Vol. 25, pp. 3875 - 3889, 2016. [DOI:10.1109/TIP.2016.2579308] [PMID]
25. [25] M. Jenadeleh, M. Ebrahimi, "BIQWS: efficient Wakeby modeling of natural scene statistics for blind image quality assessment", Multimedia Tools and Applications, Vol.76, pp. 13859-13880, 2017. [DOI:10.1007/s11042-016-3785-4]
26. [26] F. Rezaie, M. S. Helfroush, H. Danyali, "No-reference image quality assessment using local binary pattern in the wavelet domain", Multimedia Tools and Applications, Vol. 77, pp. 2529-2541, 2017. [DOI:10.1007/s11042-017-4432-4]
27. [27] L. Zhang, L. Zhang, and A. C. Bovik, "A feature-enriched completely blind image quality evaluator", IEEE Trans. Image Proc, vol. 24, pp. 2579_91, 2015. [DOI:10.1109/TIP.2015.2426416] [PMID]
28. [28] A. Mittal, G. S. Muralidhar, and J. Ghosh, "Blind image quality assessment without human training using latent quality factors", IEEE Signal Process, , vol. 19, pp.75-78, 2012. [DOI:10.1109/LSP.2011.2179293]
29. [29] W. Xue, L. Zhang, and X. Mou, "Learning without human scores for blind image quality assessment", IEEE Conf. Comput. Vis. Pattern Recognition, pp. 995-1002, 2013. [DOI:10.1109/CVPR.2013.133]
30. [30] A. Mittal, R. Soundararajan, and A. C. Bovik, "Making a completely blind image quality analyzer", IEEE SignalProcess. Lett, vol. 20, pp. 209-212, 2013. [DOI:10.1109/LSP.2012.2227726]
31. [31] P. Ye, J. Kumar, and D. Doermann, "Beyond human opinion scores: Blind image quality assessment based on synthetic scores", IEEE Int. Conf. Comput. Vis. Pattern Recognition, Columbus, pp. 4241-4248, 2014. [DOI:10.1109/CVPR.2014.540]
32. [32] L. Qiaohong, L. Weisi, F. Yuming, Z. Xinfeng, Z. Yabin, "No-reference image quality assessment based on local region statistics", Visual Communications and Image Processing, 2017.
33. [33] Q. Wu, Z. Wang, and H. Li, "A highly efficient method for blind image quality assessment", IEEE Int. Conf. Image Process., pp. 339-343,2015. [DOI:10.1109/ICIP.2015.7350816]
34. [34] M. Xiongkuo, G. Ke, Z. Guangtao, L. Jing, Y. Liu, Y. Xiaokang, C. W. Chang, "Blind Image Quality Estimation via Distortion Aggravation", IEEE Trans. Broadcasting, vol. 64, pp. 508 - 517, 2018. [DOI:10.1109/TBC.2018.2816783]
35. [35] L. Assirati, N. R. Silva, L. Berton, A. Lopes, O. M. Bruno, "Performing edge detection by Difference of Gaussians using q-Gaussian kernels", Journal of Physics, Vol. 2, 2013. [DOI:10.1088/1742-6596/490/1/012020]
36. [36] N. Inoue, K. Shinoda, "q-Gaussian mixture models for image and video semantic indexing", Journal of Visual Communication and Image Representation, Vol. 24, pp. 1450-1457, 2013. [DOI:10.1016/j.jvcir.2013.10.005]
37. [37] A. Mittal, A. K. Moorthy, and A. C. Bovik, "No-reference image quality assessment in the spatial domain", IEEE Trans. Image Process, vol. 21, pp. 4695-4908, 2012. [DOI:10.1109/TIP.2012.2214050] [PMID]
38. [38] H. R. Sheikh, M. F. Sabir, and A. C. Bovik, "A statistical evaluation of recent full reference image quality assessment algorithms", IEEE Trans. Image Process, vol. 15, pp. 3440-3451, 2006. [DOI:10.1109/TIP.2006.881959] [PMID]
39. [39] N. Ponomarenko, L. Jin, O. Ieremeiev, V. Lukin, K. Egiazarian, J. Astola, B. Vozel, K. Chehdi, M. Carli, F. Battisti, "Image database TID2013: Peculiarities, results and perspectives", Signal Processing: Image Communication, pp. 57-77, 2015. [DOI:10.1016/j.image.2014.10.009]
40. [40] T. Pouli, D. W. Cunningham, and E. Reinhard, "A survey of image statistics relevant to computer graphics", Comput. Graph. Forum, vol. 30, pp. 1761-1788, 2011. [DOI:10.1111/j.1467-8659.2011.01900.x]
41. [41] N. Ponomarenko et al, "TID2008-a database for evaluation of full-reference visual quality assessment metrics", Advances of Modern Radioelectronics, 2009.
42. [42] W. Xue, L. Zhang, X. Mou, and A. C. Bovik, "Gradient magnitude similarity deviation: A highly efficient perceptual image quality index", IEEE Trans. Image Processing, vol. 23, pp. 684-695, 2014. [DOI:10.1109/TIP.2013.2293423] [PMID]
43. [43] H. R. Sheikh, A. C. Bovik, and G. de Veciana. "An information fidelity criterion for image quality assessment using natural scene statistics", IEEE Trans. Image Processing, vol.14, pp. 2117-2128, 2005. [DOI:10.1109/TIP.2005.859389] [PMID]
44. [44] L. Zhang, D. Zhang, X. Mou, and D. Zhang. "FSIM: A feature similarity index for image quality assessment", IEEE Trans. Image Processing, vol. 20, pp. 2378-2386, 2011. [DOI:10.1109/TIP.2011.2109730] [PMID]
45. [45] Z. Wang, Q. Li. "Information content weighting for perceptual image quality assessment", IEEE Trans. Image Processing, vol. 20, pp.1185-1198, 2011. [DOI:10.1109/TIP.2010.2092435] [PMID]
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
