The article is devoted to the problem of automatic image quality assessment by a convolutional neural network when using the common TID2013 image database for training the neural network. The TID2013 database was chosen for the reason that it contains 25 base real-world images, which were distorted from these images using 24 different distortion methods and with 5 distortion levels, creating a sufficiently large database of 3000 images for training the neural network. For each image, an average expert assessment of its quality is given. All input images for the neural network are divided into two groups - the training set and the validation set. We consider two options for solving the problem of building and training a neural network and the corresponding data structure. The first option is reduced to the task of numerical regression using the numerical values of expert assessments. The second option is the task of classifying the training and validation images into 5 classes according to their quality corresponding to distortion levels. Keras and TensorFlow software tools are used to build and study the neural network. The neural network structures and relevant parameters for training each layer are presented, as well as graphs of accuracy changes for training and validation images during training.
[1] Wang, Z., Bovik, A. C. and Lu, L. “Why is image quality assessment so difficult?”, Proceedings of 2002 IEEE International Conference on Acoustics, Speech, and Signal Processing, 13-17 May 2002, Orlando, FL, USA, vol. 4, pp. IV-3313-IV-3316.
[2] Грицик, В. В. “Основні оцінки якості зображення, які сьогодні використовуються при розв’язуванні проблеми автоматичної обробки образів”, Штучний інтелект, 2017, № 1, с. 38-44.
[3] Lin, W. and Kuo, C.-C. J. “Perceptual visual quality metrics: A survey”, Journal of Visual Communication and Image Representation, May 2011, 22 (4), pp. 297-312.
[4] Soundararajan, R. and Bovik, A. “Survey of information theory in visual quality assessment”, Signal Image and Video Processing, May 2013, 7(3), pp. 391-401.
[5] Shahid, M., Rossholm, A., Lövström, B. and Zepernick, H.-J. “No-reference image and video quality assessment: a classification and review of recent approaches”, EURASIP Journal on Image and Video Processing, 2014 (1), DOI:10.1186/1687-5281-2014-40, 32 pp.
[6] Madeed, N. A., Awan, Z., and Madeed, S. A. “Image quality assessment - a survey of recent approaches”, 8th International Conference on Computer Science, Engineering and Applications, 2018, DOI:10.5121/csit.2018.80312, pp. 143-156.
[7] Ismail, T. A., Chen, S. D., Norziana, J. and Mohamad, M. A. “Improve of contrast-distorted image quality assessment based on convolutional neural networks”, International Journal of Electrical and Computer Engineering (IJECE), vol. 9, No. 6, December 2019, pp. 5604-5614.
[8] Bosse, S., Maniry, D., Müller, K.-R., Wiegand, T. and Samek, W. “Deep neural networks for no-reference and full-reference image quality assessment”, arXiv:1612.01697v2 [cs.CV] 7 Dec 2017, 14 pp.
[9] Piccini, D., Demesmaeker, R., Heerfordt, J., Yerly, J., Sopra, L. D., Masci, P. G., Schwitter, J., Van De Ville, D., Richiardi, J., Kober, T. and Stuber, M. “Deep learning to automate reference-free image quality assessment of whole-heart MR images”, Radiology: Artificial Intelligence 2020; 2(3), https://doi.org/10.1148/ryai.2020190123, 10 pp.
[10] Athar, S., Wang, Z. and Wang, Z. “Deep neural networks for blind image quality assessment: addressing the data challenge”, arXiv:2109.12161v1 [eess.IV] 24 Sep 2021, 22 pp.
[11] Cheon, M., Yoon, S.-J., Kang, B. and Lee, J. “Perceptual image quality assessment with transformers”, arXiv:2104.14730v2 [cs.CV] 5 May 2021, 10 pp.
[12] Saeed, S. U., Fu, Y., Stavrinides, V., Baum, Z. M. C., Yang, Q., Rusu, M., Fan, R. E., Sonn, G. A., Noble, J. A., Barratt, D. C. and Hu, Y. “Image quality assessment for machine learning tasks using meta-reinforcement learning”, Medical Image Analysis, 78 (2022), arXiv:2203.14258v1 [eess.IV] 27 Mar 2022, 15 pp.
[13] Dodge, S. and Karam, L. “Understanding how image quality affects deep neural networks”, arXiv:1604.04004v2 [cs.CV] 21 Apr 2016, 6 pp.
[14] Li, Y., Po, L.-M., Feng, L. and Yuan, F. “No-reference image quality assessment with deep convolutional neural networks”, 2016 IEEE International Conference on Digital Signal Processing (DSP), 16-18 October 2016, Beijing, China, DOI: 10.1109/ICDSP.2016.7868646, pp. 685-689.
[15] Simonyan, K. and Zisserman, A. “Very deep convolutional networks for large-scale image recognition”, arXiv:1409.1556v6 [cs.CV] 10 Apr 2015, 14 pp.
[16] Bianco, S., Celona, L., Napoletano, P. and Schettini, R. “On the use of deep learning for blind image quality assessment”, arXiv:1602.05531v5 [cs.CV] 4 Apr 2017, 8 pp.
[17] Gu, K., Zhai, G., Yang, X. and Zhang, W. “Deep learning network for blind image quality assessment”, 2014 IEEE International Conference on Image Processing (ICIP), 27-30 October 2014, DOI: 10.1109/ICIP.2014.7025102, Paris, France, pp. 511-515.
[18] Ponomarenko, N., Jin, L., Ieremeiev, O., Lukin, V., Egiazarian, K., Astola, J., Vozel, B., Chehdi, K., Carli, M., Battisti, F. and Kuo, C.-C. J. “Image database TID2013: Peculiarities, results and perspectives”, Signal Processing: Image Communication, 30 (2015), pp. 57-77.
[19] Ponomarenko, N., Ieremeiev, O., Lukin, V., Jin, L., Egiazarian, K., Astola, J., Vozel, B., Chehdi, K., Carli, M., Battisti, F. and Kuo C.-C. J. “A new color image database TID2013: Innovations and results”, Proceedings of ACIVS, Poznan, Poland, Oct. 2013, pp. 402-413.
[20] Chollet, F. Deep Learning with Python, 2nd ed., Manning Publications Co., 2021, 478 pp.