Abstract
In recent years, the image deblurring task has attracted more and more researchers’ attention. Many researchers are devoted to eliminating motion blur by using a “coarse-to-fine” architecture. This architecture can effectively eliminate motion blur caused by a simple relative displacement. But when using the architecture directly for the face image deblurring task, there would exist some problems. For example, complex network structures make the model difficult to be trained, and a large number of parameters need to be calculated would result in expensive runtime. Since details of images can’t be completely restored, the quality of images will deteriorate, and deblurring visual effect is poor, so in order to solve these issues, combining the “coarse-to-fine” architecture with the “stacked” architecture, we propose a new network architecture-“Stack-based Scale-recurrent Network” for the face image deblurring task. The ConvLSTM network is employed to build a model. We achieves the purpose of improving the visual clarity and the quality of restored face images. And we use the improved encoder-decoder to enhance network performance. Compared with the other deblurring methods, our method can restore clear face images with higher-quality and clearer vision.
Similar content being viewed by others
References
Wang Z, Yao Z, Wang Q (2017) Improved scheme of estimating motion blur parameters for image restoration[J]. Digit Sign Process 65:11–18
Hirsch M, Schuler C J, Harmeling S et al (2012) Fast removal of non-uniform camera shake[C]. In: IEEE International conference on computer vision. IEEE
Harmeling S, Hirsch M, Schlkopf B (2010) Space-variant single-image blind deconvolution for removing camera shake[C]. Adv Neural Inf Process Syst 23:829–837
Perrone D, Favaro P (2016) A logarithmic image prior for blind deconvolution [J]. Int J Comput Vis 117(2):159–172
Jiang X, Yao H, Zhao S (2017) Text image deblurring via two-tone prior [J]. Neurocomputing 242:1–14
Xu Z, Chen H, Li Z (2021) Fast blind deconvolution using a deeper sparse patch-wise maximum gradient prior[J]. Signal Process Image Commun 90:1114–1122
Pan J, Sun D, Pfifister H et al (2016) Blind image deblurring using dark channel prior[C]. In: IEEE Conference on computer vision and pattern recognition. IEEE, pp. 1628–1636
Yi O (2019) Total variation constraint GAN for dynamic scene deblurring[J]. Image Vis Comput 88:113–119
Kja B, Ying SC, Ql C et al (2021) Image restoration using overlapping group sparsity on hyper-Laplacian prior of image gradient[J]. Neurocomputing 420:57–69
Zhou K, Zhuang P, Xiong J et al (2020) Blind image deblurring with joint extreme channels and L0-regularized intensity and gradient priors[C]. In: IEEE International conference on image processing (ICIP). IEEE
Ledig C, Theis L, Huszar F, Caballero J, Cunningham A, Acosta A, Aitken A, Tejani A, Totz J, Wang Z, Shi W (2017) Photo-realistic single image super–resolution using a generative adversarial network. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, pp 105–114
Yda B, Zd C, Shuai YC et al (2020) Kernel-attended residual network for single image super-resolution - ScienceDirect[J]. Knowledge-Based Syst 213:52–60
Yu J, Tan M, Zhang H et al (2019) Hierarchical deep click feature prediction for fine-grained image recognition[J]. IEEE Trans Pattern Anal Mach Intell 99:1
Isola P, Zhu J Y, Zhou T et al (2016) Image-to-image translation with conditional adversarial networks[C]. In: IEEE Conference on computer vision and pattern recognition. IEEE
Yu J, Kuang Z, Zhang B et al (2018) Leveraging content sensitiveness and user trustworthiness to recommend fine-grained privacy settings for social image sharing[J]. IEEE Trans Inf Forensics Secur 13:1317–1332
Li Y, Huang JB, Ahuja N et al (2016) Deep joint image filtering[J]. In: European conference on computer vision. pp. 771–779
Khan AT, Cao X, Li S et al (2021) Quantum beetle antennae search: a novel technique for the constrained portfolio optimization problem[J]. Sci China Inf Sci 64(5):152204
Atk A, Shuai LA, Xc B (2021) Control framework for cooperative robots in smart home using bio-inspired neural network - ScienceDirect[J]. Measurement 167:108253
Gampala V, Kumar M, Sushama C, Sehar E, Raj FI (2020) Deep learning based image processing approaches for image deblurring. Mater Today Proc 2020:601–609
Zhang J, Pan J, Ren J et al (2018) Dynamic scene deblurring using spatially variant recurrent neural networks[C]. In: IEEE/CVF Conference on computer vision and pattern recognition. IEEE
Huang L, Xia Y (2019) Joint blur Kernel estimation and CNN for blind image restoration[J]. Neurocomputing 396:562–570
Singhal J, DeblurRL Narang P (2021) Image deblurring with deep reinforcement learning[M]. Electronic Industry Press, Beijing
Wang M, Hou S, Li H et al (2019) Generative image deblurring based on multi-scaled residual adversary network driven by composed prior-posterior loss[J]. J Vis Commun Image Represent 65:1621–1629
Xu L, Ren JS, Liu C et al (2014) Deep convolutional neural network for image deconvolution[C]. NIPS 27:1790–1798
Gong D, Yang J, Liu L et al (2017) From motion blur to motion flflow: a deep learning solution for removing heterogeneous motion blur[C]. In: CVPR
Noroozi M, Chandramouli P, Favaro P (2017) Motion deblurring in the wild[C]. In: German conference on pattern recognition
Sun J, Cao W, Xu Z et al (2015) Learning a convolutional neural network for non-uniform motion blur removal[J]. In: IEEE Conference on computer vision and pattern recognition pp. 235–243
Nah S, Kim TH, Lee KM (2017) Deep multi-scale convolutional neural network for dynamic scene deblurring. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, pp 257–265
Chakrabarti A (2016) A neural approach to blind motion deblurring[C]. In: European conference on computer vision. Springer, Cham, pp 221–235
Dong G, Jie Y, Liu L et al (2017) From Motion blur to motion flow: a deep learning solution for removing heterogeneous motion blur[C]. In: IEEE Conference on computer vision and pattern recognition. IEEE
Seungjun Nah, Tae Hyun Kim, Kyoung Mu Lee (2017) Deep multi-scale convolutional neural network for dynamic scene deblurring[C]. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 257–265
Mao XJ, Shen C, Yang YB (2016) Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connections[J]. In: Conference on neural information processing systems pp. 2802–2810
Hongguang Zhang, Yuchao Dai, Hongdong Li et al (2019) Deep stacked hierarchical multi-patch network for image deblurring[C]. In: CVPR, pp.5979-5986
Kupyn O, Budzan V, Mykhailych M et al (2018) DeblurGAN: Blind motion deblurring using conditional adversarial networks[C]. In: IEEE/CVF Conference on computer vision and pattern recognition. IEEE
Yu J, Zhu C, Zhang J et al (2019) Spatial pyramid-enhanced NetVLAD with weighted triplet loss for place recognition[J]. IEEE Trans Neural Netw Learn Syst 31:1–14
Eigen D, Fergus R (2014) Predicting depth, surface normals and semantic labels with a common multi-scale convolutional architecture[C]. In: IEEE. IEEE
Mathieu MF, Couprie C, Cun Y (2018) Deep multi-scale video prediction. US20180137389[P]
Dosovitskiy A, Fischer P, Ilg E et al (2015) Flownet: Learning optical flflow with convolutional networks. In: CVPR, pp. 2758–2766
Khan AH, Li S, Luo X (2020) Obstacle avoidance and tracking control of redundant robotic manipulator: an RNN-based metaheuristic approach[J]. IEEE Trans Ind Inform 16(7):4670–4680
Khan AH, Li S, Cao X (2021) Tracking control of redundant manipulator under active remote center-of-motion constraints: an RNN-based metaheuristic approach[J]. Sci China Inf Sci 64(3):1–18
Ahk A, Shuai LB, Dc C et al (2020) Tracking control of redundant mobile manipulator: an RNN based metaheuristic approach[J]. Neurocomputing 400:272–284
Tao X, Gao H, Shen X, Wang J, Jia J (2018) Scale-recurrent network for deep image deblurring[C]. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 8174–8182
Liu Z, Yeh R, Tang X et al (2017) Video frame synthesis using deep voxel flow[J]. In: IEEE
Su S, Delbracio M, Wang J et al. (2016) Deep video deblurring [J]
Tao X, Gao H, Liao R et al (2017) Detail-revealing deep video super-resolution. In: IEEE Computer society
Xu N, Price B, Cohen S et al (2017) Deep Image Matting[C]. In: IEEE Conference on computer vision and pattern recognition. IEEE Computer society, pp. 311–320
He K, Zhang X, Ren S et al (2016) Deep Residual Learning for Image Recognition[C]. In: 2016 IEEE Conference on computer vision and pattern recognition (CVPR). IEEE, 2016
Kingma D, Adam Ba J (2014) A method for stochastic optimization[J]. Computer. Science
Acknowledgements
This work was partly supported in part by the National Natural Science Foundation of China (No. 61871464 and 61836002), the Fujian Provincial Natural Science Foundation of China (No. 2018J01573 and 2019J01854), the Open Project Program of State Key Laboratory of Virtual Reality Technology and Systems, Beihang University (No. VRLAB2021C01), Distinguished Young Scientific Research Talents Plan in Universities of Fujian Province, the Program for New Century Excellent Talents in University of Fujian Province and Xiamen Youth Innovation Fund Project (No. 3502Z20206072).
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
Wu, Y., Hong, C., Zhang, X. et al. Stack-based Scale-recurrent Network for Face Image Deblurring. Neural Process Lett 53, 4419–4436 (2021). https://doi.org/10.1007/s11063-021-10604-9
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11063-021-10604-9