Abstract
Technological advancements continue to expand the communications industry’s potential. Images, which are an important component in strengthening communication, are widely available. Therefore, image quality assessment (IQA) is critical in improving content delivered to end users. Convolutional neural networks (CNNs) used in IQA face two common challenges. One issue is that these methods fail to provide the best representation of the image. The other issue is that the models have a large number of parameters, which easily leads to overfitting. To address these issues, the dense convolution network (DSC-Net), a deep learning model with fewer parameters, is proposed for no-reference image quality assessment (NR-IQA). Moreover, it is obvious that the use of multimodal data for deep learning has improved the performance of applications. As a result, multimodal dense convolution network (MDSC-Net) fuses the texture features extracted using the gray-level co-occurrence matrix (GLCM) method and spatial features extracted using DSC-Net and predicts the image quality. The performance of the proposed framework on the benchmark synthetic datasets LIVE, TID2013, and KADID-10k demonstrates that the MDSC-Net approach achieves good performance over state-of-the-art methods for the NR-IQA task.
摘要
科技进步不断扩大通信行业的潜力. 图像在加强交流中发挥着重要作用, 已被广泛应用. 因此, 图像质量评估(IQA)对优化传递给终端用户的内容至关重要. 在IQA中使用卷积神经网络面临两个常见难题. 一是这些方法难以提供图像最佳表示, 另一个问题是模型具有大量参数, 容易导致过拟合. 为解决这些问题, 提出一种参数更少的深度学习模型——密集卷积网络(DSC-Net), 用于无参考图像质量评估(NR-IQA). 此外, 将多模态数据用于深度学习明显改进各种应用的性能. 多模态密集卷积网络(MDSC-Net)融合了灰度共生矩阵(GLCM)方法提取的纹理特征和DSC-Net方法提取的空间特征, 并对图像质量进行预测. 所提框架在基准合成数据集LIVE、 TID2013和KADID-10k的性能表明, MDSC-Net方法在NR-IQA任务中表现出良好性能, 超过了当前最先进的方法.
This is a preview of subscription content, log in via an institution to check access.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Bianco S, Celona L, Napoletano P, et al., 2018. On the use of deep learning for blind image quality assessment. Signal Image Video Process, 12(2):355–362. https://doi.org/10.1007/s11760-017-1166-8
Bosse S, Maniry D, Wiegand T, et al., 2016. A deep neural network for image quality assessment. Proc IEEE Int Conf on Image Processing, p.3773–3777. https://doi.org/10.1109/ICIP.2016.7533065
Bosse S, Maniry D, Müller KR, et al., 2018. Deep neural networks for no-reference and full-reference image quality assessment. IEEE Trans Image Process, 27(1):206–219. https://doi.org/10.1109/TIP.2017.2760518
Chockalingam N, Murugan B, 2023. Hierarchical patch selection: an improved patch sampling for no reference image quality assessment. IEEE Trans Artif Intell, early access. https://doi.org/10.1109/TAI.2023.3262623
Cheng ZX, Takeuchi M, Katto J, 2017. A pre-saliency map based blind image quality assessment via convolutional neural networks. Proc IEEE Int Symp on Multimedia, p.77–82. https://doi.org/10.1109/ISM.2017.21
Deng J, Dong W, Socher R, et al., 2009. ImageNet: a large-scale hierarchical image database. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.248–255. https://doi.org/10.1109/CVPR.2009.5206848
Ding GG, Chen WS, Zhao SC, et al., 2018. Real-time scalable visual tracking via quadrangle kernelized correlation filters. IEEE Trans Intell Transp Syst, 19(1):140–150. https://doi.org/10.1109/CVPR.2009.5206848
Ding GG, Guo YC, Chen K, et al., 2019. DECODE: deep confidence network for robust image classification. IEEE Trans Image Process, 28(8):3752–3765. https://doi.org/10.1109/TIP.2019.2902115
Gu K, Tao DC, Qiao JF, et al., 2018. Learning a no-reference quality assessment model of enhanced images with big data. IEEE Trans Neur Netw Learn Syst, 29(4):1301–1313. https://doi.org/10.1109/TNNLS.2017.2649101
Gu K, Xia ZF, Qiao JF, et al., 2020. Deep dual-channel neural network for image-based smoke detection. IEEE Trans Multimed, 22(2):311–323. https://doi.org/10.1109/TMM.2019.2929009
Gu K, Zhang YH, Qiao JF, 2021a. Ensemble meta-learning for few-shot soot density recognition. IEEE Trans Industr Inform, 17(3):2261–2270. https://doi.org/10.1109/TII.2020.2991208
Gu K, Liu HY, Xia ZF, et al., 2021b. PM2.5 monitoring: use information abundance measurement and wide and deep learning. IEEE Trans Neur Netw Learn Syst, 32(10):4278–4290. https://doi.org/10.1109/TNNLS.2021.3105394
He KM, Zhang XY, Ren SQ, et al., 2016. Deep residual learning for image recognition. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.770–778. https://doi.org/10.1109/CVPR.2016.90
Huang G, Liu Z, Van Der Maaten L, et al., 2017. Densely connected convolutional networks. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.4700–4708. https://doi.org/10.1109/CVPR.2017.243
Kang L, Ye P, Li Y, et al., 2014. Convolutional neural networks for no-reference image quality assessment. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.1733–1740. https://doi.org/10.1109/CVPR.2014.224
Kang L, Ye P, Li Y, et al., 2015. Simultaneous estimation of image quality and distortion via multi-task convolutional neural networks. Proc IEEE Int Conf on Image Processing, p.2791–2795. https://doi.org/10.1109/ICIP.2015.7351311
Kim J, Lee S, 2017. Fully deep blind image quality predictor. IEEE J Sel Top Signal Process, 11(1):206–220. https://doi.org/10.1109/JSTSP.2016.2639328
Krizhevsky A, Sutskever I, Hinton GE, 2012. ImageNet classification with deep convolutional neural networks. Proc 25th Int Conf on Neural Information Processing Systems, p.1097–1105.
Li QH, Lin WS, Xu JT, et al., 2016. Blind image quality assessment using statistical structural and luminance features. IEEE Trans Multimed, 18(12):2457–2469. https://doi.org/10.1109/TMM.2016.2601028
Li ZC, Tang JH, Mei T, 2019. Deep collaborative embedding for social image understanding. IEEE Trans Patt Anal Mach Intell, 41(9):2070–2083. https://doi.org/10.1109/TPAMI.2018.2852750
Lin HH, Hosu V, Saupe D, 2019. KADID-10k: a large-scale artificially distorted IQA data-base. Proc 11th Int Conf on Quality of Multimedia Experience, p.1–3. https://doi.org/10.1109/QoMEX.2019.8743252
Lin TY, RoyChowdhury A, Maji S, 2015. Bilinear CNN models for fine-grained visual recognition. Proc IEEE Int Conf on Computer Vision, p.1449–1457. https://doi.org/10.1109/ICCV.2015.170
Liu LX, Liu B, Huang H, et al., 2014. No-reference image quality assessment based on spatial and spectral entropies. Signal Process Image Commun, 29(8):856–863. https://doi.org/10.1016/j.image.2014.06.006
Liu XL, Van De Weijer J, Bagdanov AD, 2017. RankIQA: learning from rankings for no-reference image quality assessment. Proc IEEE Int Conf on Computer Vision, p.1040–1049. https://doi.org/10.1109/ICCV.2017.118
Lu ZK, Lin W, Yang X, et al., 2005. Modeling visual attention’s modulatory aftereffects on visual sensitivity and quality evaluation. IEEE Trans Image Process, 14(11):1928–1942. https://doi.org/10.1109/TIP.2005.854478
Ma JP, Wu JJ, Li LD, et al., 2021. Blind image quality assessment with active inference. IEEE Trans Image Process, 30:3650–3663. https://doi.org/10.1109/TIP.2021.3064195
Ma KD, Liu WT, Zhang K, et al., 2018. End-to-end blind image quality assessment using deep neural networks. IEEE Trans Image Process, 27(3):1202–1213. https://doi.org/10.1109/TIP.2017.2774045
Mittal A, Moorthy AK, Bovik AC, 2012. No-reference image quality assessment in the spatial domain. IEEE Trans Image Process, 21(12):4695–4708. https://doi.org/10.1109/TIP.2012.2214050
Moorthy AK, Bovik AC, 2011. Blind image quality assessment: from natural scene statistics to perceptual quality. IEEE Trans Image Process, 20(12):3350–3364. https://doi.org/10.1109/TIP.2011.2147325
Pan ZQ, Yuan F, Lei JJ, et al., 2022. VCRNet: visual compensation restoration network for no-reference image quality assessment. IEEE Trans Image Process, 31:1613–1627. https://doi.org/10.1109/TIP.2022.3144892
Po LM, Liu MY, Yuen WYF, et al., 2019. A novel patch variance biased convolutional neural network for no-reference image quality assessment. IEEE Trans Circ Syst Video Technol, 29(4):1223–1229. https://doi.org/10.1109/TCSVT.2019.2891159
Ponomarenko N, Jin LN, Ieremeiev O, et al., 2015. Image database TID2013: peculiarities, results and perspectives. Signal Process Image Commun, 30:57–77. https://doi.org/10.1016/j.image.2014.10.009
Qiu ZF, Yao T, Mei T, 2018. Learning deep spatio-temporal dependence for semantic video segmentation. IEEE Trans Multimed, 20(4):939–949. https://doi.org/10.1109/tmm.2017.2759504
Ren HY, Chen DQ, Wang YZ, 2018. RAN4IQA: restorative adversarial nets for no-reference image quality assessment. Proc 32nd AAAI Conf on Artificial Intelligence, p.7308–7314. https://doi.org/10.1609/aaai.v32i1.12258
Saad MA, Bovik AC, Charrier C, 2012. Blind image quality assessment: a natural scene statistics approach in the DCT domain. IEEE Trans Image Process, 21(8):3339–3352. https://doi.org/10.1109/TIP.2012.2191563
Sheikh HR, 2003. Image and Video Quality Assessment Research at Live. http://liveeceutexas.edu/research/quality [Accessed on Oct. 30, 2022].
Sheikh HR, Bovik AC, Cormack L, 2003. Blind quality assessment of JEPG2000 compressed images using natural scene statistics. Proc 37th Asilomar Conf on Signals, Systems & Computers, p.1403–1407. https://doi.org/10.1109/ACSSC.2003.1292217
Simonyan K, Zisserman A, 2014. Very deep convolutional networks for large-scale image recognition. https://arxiv.org/abs/1409.1556
Song GH, Jin XG, Chen GL, et al., 2016. Two-level hierarchical feature learning for image classification. Front Inform Technol Electron Eng, 17(9):897–906. https://doi.org/10.1631/FITEE.1500346
Tang HX, Joshi N, Kapoor A, 2011. Learning a blind measure of perceptual image quality. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.305–312. https://doi.org/10.1109/CVPR.2011.5995446
Wang Z, Shang XL, 2006. Spatial pooling strategies for perceptual image quality assessment. Proc Int Conf on Image Processing, p.2945–2948. https://doi.org/10.1109/ICIP.2006.313136
Wu JJ, Zhang M, Li LD, et al., 2019. No-reference image quality assessment with visual pattern degradation. Inform Sci, 504:487–500. https://doi.org/10.1016/j.ins.2019.07.061
Xu JT, Ye P, Li QH, et al., 2016. Blind image quality assessment based on high order statistics aggregation. IEEE Trans Image Process, 25(9):4444–4457. https://doi.org/10.1109/TIP.2016.2585880
Yang GY, Ding XY, Huang T, et al., 2020. Explicit-implicit dual stream network for image quality assessment. EURASIP J Image Video Process, 2020(1):48. https://doi.org/10.1186/s13640-020-00538-y
Ye P, Kumar J, Kang L, et al., 2012. Unsupervised feature learning framework for no-reference image quality assessment. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.1098–1105. https://doi.org/10.1109/CVPR.2012.6247789
Zhang P, Zhou WG, Wu L, et al., 2015. SOM: semantic obviousness metric for image quality assessment. Proc IEEE Conf on Computer Vision and Pattern Recognition, p.2394–2402. https://doi.org/10.1109/CVPR.2015.7298853
Zhang SQ, Zhang SL, Huang TJ, et al., 2018. Speech emotion recognition using deep convolutional neural network and discriminant temporal pyramid matching. IEEE Trans Multimed, 20(6):1576–1590. https://doi.org/10.1109/TMM.2017.2766843
Zhang WX, Ma KD, Yan J, et al., 2020. Blind image quality assessment using a deep bilinear convolutional neural network. IEEE Trans Circ Syst Video Technol, 30(1):36–47. https://doi.org/10.1109/TCSVT.2018.2886771
Zhang WX, Ma KD, Zhai GT, et al., 2021. Uncertainty-aware blind image quality assessment in the laboratory and wild. IEEE Trans Image Process, 30:3474–3486. https://doi.org/10.1109/TIP.2021.3061932
Zhou ZH, Lu W, Yang JC, et al., 2020. No-reference image quality assessment based on neighborhood cooccurrence matrix. Signal Process Image Commun, 81:115680. https://doi.org/10.1016/j.image.2019.115680
Author information
Authors and Affiliations
Contributions
Nandhini CHOCKALINGAM designed the research, processed the data, and drafted the paper. Brindha MURUGAN helped organize and revise the paper. Nandhini CHOCKALINGAM and Brindha MURUGAN finalized the paper.
Corresponding author
Ethics declarations
Nandhini CHOCKALINGAM and Brindha MURUGAN declare that they have no conflict of interest.
Additional information
List of supplementary materials
1 Description of the dataset
2 Details of evaluation metrics
3 Local contrast normalization (LCN)
4 Importance of GLCM features for influential performance
5 Visualization of distortion specific prediction
6 Effect of patch size
7 Performance of pre-trained DenseNet
8 Visualization of patchwise training strategy
9 Convergence analysis
Supplementary materials for
Rights and permissions
About this article
Cite this article
Chockalingam, N., Murugan, B. A multimodal dense convolution network for blind image quality assessment. Front Inform Technol Electron Eng 24, 1601–1615 (2023). https://doi.org/10.1631/FITEE.2200534
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1631/FITEE.2200534
Key words
- No-reference image quality assessment (NR-IQA)
- Blind image quality assessment
- Multimodal dense convolution network (MDSC-Net)
- Deep learning
- Visual quality
- Perceptual quality