Abstract
Recently, single image super-resolution (SISR) based on convolutional neural networks (CNNs) has represented great progress. However, due to the huge number of parameters, these models cannot work well in many real-world applications, most of which fail to exploit the multi-scale features and the hierarchical features for lightweight and accurate image SR. In this paper, a lightweight multi-scale aggregated residual attention network (MARAN) is proposed by exploring multi-scale contextual information and multi-level features. The network consists of shallow feature extraction, recursively stacked multiple multi-scale aggregated residual attention groups (MARAGs), multi-level feature fusion block (MLFFB), and reconstruction part. Specifically, the MARAGs produce the hierarchical multi-scale deep features, the MLFFB effectively fuses the hierarchical features with multi-scale aggregated residual attention. Each MARAG is composed of cascaded multi-scale aggregated residual attention blocks (MARABs) and each MARAB contains a multi-scale aggregated unit and a dual-attention unit. The multi-scale aggregated unit expands group convolution with cross-path connection. The dual-attention unit can adaptively modulate region-based information and channel-wise features. Qualitative and quantitative experiments on four benchmark datasets demonstrate that the proposed MARAN achieves better performance against state-of-the-art methods with fewer parameters.
Similar content being viewed by others
References
Ahn N, Kang B, Sohn K (2018) Fast, accurate, and lightweight super-resolution with cascading residual network. In: The European conference on computer vision (ECCV), Munich, Germany, pp 252– 268
Aitken A, Ledig C, Theis L, Caballero J, Wang Z, Shi W (2017) Checkerboard artifact free sub-pixel convolution: A note on sub-pixel convolution, resize convolution and convolution resize, arXiv:1707.02937
Anagun Y, Isik S, Seke E (2019) SRLIbrary: comparing different loss functions for super-resolution over various convolutional architectures. J Vis Commun Image Represent 61:178–187
Arbel’aez P, Maire M, Fowlkes C, Malik J (2011) Contour detection and hierarchical image segmentation. IEEE Trans Pattern Anal Mach Intell 33(5):898–916
Bevilacqua M, Roumy A, Guillemot C, Alberi-Morel ML (2012) Low-complexity single image super-resolution based on nonnegative neighbor embedding. In: British machine vision conference (BMVC), BMVA press surrey, UK, pp 1–10
Brownlee J (2016) Overfitting and underfitting with machine learning algorithms. Machine Learning Mastery 21
Cao F, Chen B (2019) New architecture of deep recursive convolution networks for super-resolution. Knowl-Based Syst 178:98–110
Chen L, Zhang H, Xiao J, Nie L, Shao J, Liu W, Chua T-S (2017) Sca-cnn: Spatial and channel-wise attention in convolutional networks for image captioning. In: IEEE conference on computer vision and pattern recognition (CVPR), Honolulu, USA, pp 5659–5667
Dong C, Loy CC, He K, Tang X (2014) Learning a deep convolutional network for image super-resolution. In: The European conference on computer vision (ECCV), pp 184–199
Dong C, Loy CC, Tang X (2016) Accelerating the super-resolution convolutional neural network. In: The European conference on computer vision (ECCV), The Netherlands, pp 391–407
Gao S, Cheng MM, Zhao K, Zhang XY, Yang MH, Torr PH (2019) Res2net: A new multi-scale backbone architecture. IEEE Trans Pattern Anal Mach Intell 1–1
He K, Zhang X, Ren S. h., Sun J (2016) Deep residual learning for image recognition. In: IEEE conference on computer vision and pattern recognition (CVPR), Las Vegas, USA, pp 770–778
Hou Q, Cheng M-M, Hu X, Borji A, Tu Z, Torr P (2017) Deeply supervised salient object detection with short connections. In: IEEE conference on computer vision and pattern recognition (CVPR). Honolulu, USA, pp 5300–5309
Howard AG, Zhu M, Chen B, Kalenichenko D, Wang W, Weyand T, Andreetto M, Adam H (2017) Mobilenets: Efficient convolutional neural networks for mobile vision applications, arXiv:1704.04861
Hu Y, Li J, Huang Y, Gao X (2019) Channel-wise and spatial feature modulation network for single image super-resolution. IEEE Transactions on Circuits and Systems for Video Technology
Hu J, Shen L, Sun G (2018) Squeeze-and-excitation networks. In: IEEE conference on computer vision and pattern recognition (CVPR), Salt Lake City, USA, pp 7132–7141
Huang J-B, Singh A, Ahuja N (2015) Single image super-resolution from transformed self-exemplars. In: The IEEE conference on computer vision and pattern recognition (CVPR), Boston, USA, pp 5197–5206
Huang J-J, Siu W-C, Liu T-R (2015) Fast image interpolation via random forests. IEEE Trans Image Process 24(10):3232–3245
Hui Z, Wang X, Gao X (2018) Fast and accurate single image super-resolution via information distillation network. In: IEEE conference on computer vision and pattern recognition (CVPR), Salt Lake City, USA, pp 723–731
Jiang K, Wang Z, Yi P, Jiang J (2018) A progressively enhanced network for video satellite imagery super-resolution. IEEE Signal Process Lett 25 (11):1630–1634
Jing P, Guan W, Bai X, Guo H, Su Y (2020) Single image super-resolution via low-rank tensor representation and hierarchical dictionary learning. Multimed Tools Appl 1–19
Kim J, Kwon Lee J, Lee K. M. u. (2016) Deeply-recursive convolutional network for image super-resolution. In: IEEE conference on computer vision and pattern recognition (CVPR), Las Vegas, USA, pp 1637–1645
Kim J, Lee JK, Lee KM (2016) Accurate image super-resolution using very deep convolutional networks. In: IEEE conference on computer vision and pattern recognition (CVPR), Las Vegas, USA, pp 1646–1654
Kingma DP, Ba J (2015) Adam: A method for stochastic optimization. In: International conference for learning representations (ICLR), San Diego, USA, pp 1–13
Lai WS, Huang JB, Ahuja N, Yang MH (2017) Deep laplacian pyramid networks for fast and accurate super-resolution. In: IEEE conference on computer vision and pattern recognition (CVPR), Honolulu, USA, pp 624–632
Ledig C, Theis L, Husz’ar F, Caballero J (2017) Photo-realistic single image super-resolution using a generative adversarial network. In: IEEE conference on computer vision and pattern recognition (CVPR), Honolulu, USA, pp 4681–4690
Lim B, Son S, Kim H, Nah S, Lee KM (2017) Enhanced deep residual networks for single image super-resolution. In: IEEE conference on computer vision and pattern recognition workshops (CVPRW). Honolulu, USA, pp 136–144
Lim B, Son S, Kim H, Nah S, Lee KM (2017) Enhanced deep residual networks for single image super-resolution. In: IEEE conference on computer vision and pattern recognition (CVPR), Honolulu, USA, pp 136–144
Lin F, Fookes C, Chandran V, Sridharan S (2007) Super-resolved faces for improved face recognition from surveillance video. In: The international conference on biometrics. Springer, Berlin, pp 1–10
Lu P, Barazzetti L, Chandran V, Gavaghan K, Weber S, Gerber N, Reyes M (2018) Highly accurate facial nerve segmentation refinement from CBCT/CT imaging using a super-resolution classification approach. IEEE Trans Biomed Eng 65(1):178–188
Pang S, Chen Z, Yin F (2021) Convolutional neural network based sub-pixel line-edged angle detection with applications in measurement. IEEE Sensors J 21(7):9314–9322
Paszke A, Gross S, Chintala S, Chanan G, Yang E, DeVito Z, Lin Z, Desmaison A, Antiga L, Lerer A (2017) Automatic differentiation in PyTorch. Advances in Neural Information Processing Systems Autodiff Workshop 1–4
Qiao J, Song H, Zhang K, Zhang X, Liu Q (2019) Image super-resolution using conditional generative adversarial network. IET Image Process 13 (14):2673–2679
Romano Y, Protter M, Elad M (2014) Single image interpolation via adaptive nonlocal sparsity-based modeling. IEEE Trans Image Process 23 (7):3085–3098
Shi W, Caballero J, Huszar F, Totz J, Aitken AP, Bishop R, Rueckert D, Wang Z (2016) Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network. In: IEEE conference on computer vision and pattern recognition (CVPR), Las Vegas, NV, USA, pp 1874– 1883
Shi W, Caballero J, Ledig C, Zhuang X, Bai W, Bhatia K, de Marvao AMSM, Dawes T, ORegan D, Rueckert D (2013) Cardiac image super-resolution with global correspondence using multi-atlas patchmatch. In: The international conference on medical image computing and computer-assisted intervention. Berlin, Heidelberg, pp 9–16
Tai Y, Yang J, Liu X (2017) Image super-resolution via deep recursive residual network. In: IEEE conference on computer vision and pattern recognition (CVPR), Honolulu, USA, pp 3147–3155
Tai Y, Yang J, Liu X, Xu C (2017) Memnet: A persistent memory network for image restoration. In: IEEE international conference on computer vision (ICCV), Venice, Italy, pp 4539–4547
Tang Z, Li S, Luo L, Fu M, Peng H, Zhou Q (2019) Image super-resolution via simplified dense network with non-degenerate layers. IEEE Access 24775–24787
Timofte R, Agustsson E, Van Gool L, Yang M. -H., Zhang L (2017) Ntire 2017 challenge on single image super-resolution: methods and results. In: IEEE conference on computer vision and pattern recognition workshops (CVPRW), Honolulu, USA, pp 114–125
Wang Z, Chen J, Hoi SC (2020) Deep learning for image super-resolution: a survey. IEEE Trans Pattern Anal Mach Intell 1–1
Wang F, Jiang M, Qian C, Yang S, Li C, Zhang H, Wang X, Tang X (2017) Residual attention network for image classification. In: IEEE conference on computer vision and pattern recognition (CVPR), Honolulu, USA, pp 6450–6458
Xue S, Qiu W, Liu F, Jin X (2020) Wavelet-based residual attention network for image super-resolution. Neurocomputing 382:116–126
Yang W, Wang W, Zhang X, Sun S, Liao Q (2019) Lightweight feature fusion network for single image super-resolution. IEEE Signal Process Lett 26(4):538–542
Yang A, Yang B, Ji Z, Pang Y, Shao L (2020) Lightweight group convolutional network for single image super-resolution. Inf Sci 516:220–233
Yang X, Zhang Y, Guo Y, Zhou D (2021) An image super-resolution deep learning network based on multi-level feature extraction module. Multimed Tools Appl 80(5):7063–7075
Zeyde R, Elad M, Protter M (2010) On single image scale-up using sparse representations. In: International conference on curves and surfaces, Heidelberg, Berlin, pp 711–730
Zhang K, Gao X, Tao D, Li X (2012) Single image super-resolution with non-local means and steering kernel regression. Trans Image Process 21 (11):4544–4556
Zhang Y, Li K, Li K, Wang L, Zhong B, Fu Y (2018) Image super-resolution using very deep residual channel attention networks. In: European conference on computer vision (ECCV), Munich, Germany, pp 286–301
Zhang Y, Li K, Li K, Zhong B, Fu Y (2019) Residual non-local attention networks for image restoration, arXiv:1903.10082
Zhang W, Liu Y, Dong C, Qiao Y (2019) Ranksrgan: Generative adversarial networks with ranker for image super-resolution. In: IEEE international conference on computer vision (ICCV), Seoul, Korea, pp 3096–3105
Zhang Y, Tian Y, Kong Y, Zhong B, Fu Y (2018) Residual dense network for image super-resolution. In: IEEE conference on computer vision and pattern recognition (CVPR), Salt Lake City, USA, pp 2472–2481
Zhang S, Yuan Q, Li J, Sun J, Zhang X (2020) Scene-adaptive remote sensing image super-resolution using a multiscale attention network. IEEE Transactions on Geoscience and Remote Sensing
Zhao L, Bai H, Liang J, Zeng B, Wang A, Zhao Y (2019) Simultaneous color-depth super-resolution with conditional generative adversarial networks. Pattern Recogn 88:356–369
Zhou L, Wang Z, Luo Y, Xiong Z (2019) Separability and compactness network for image recognition and super-resolution. IEEE Trans Neural Netw Learn Syst 30(11):3275–3286
Zhu L, Zhan S, Zhang HJN (2019) Stacked U-shape networks with channel-wise attention for image super-resolution. Neurocomputing 345:58–66
Zou WWW, Yuen PC (2012) Very low resolution face recognition problem. IEEE Trans Image Process 21(1):327–340
Acknowledgments
This work was supported by Natural Science Foundations of China (Nos.61771091, 61871066), National High Technology Research and Development Program (863 Program) of China (No. 2015AA016306), Natural Science Foundation of Liaoning Province of China (No. 20170540159), and Fundamental Research Fund for the Central Universities of China (No.DUT17LAB04).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Pang, S., Chen, Z. & Yin, F. Lightweight multi-scale aggregated residual attention networks for image super-resolution. Multimed Tools Appl 81, 4797–4819 (2022). https://doi.org/10.1007/s11042-021-11138-x
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-021-11138-x