Scale-Recurrent Network for Deep Image Deblurring

By Xin Tao, Hongyun Gao, Yi Wang, Xiaoyong Shen, Jue Wang, Jiaya Jia. 2018

1,306 citations recorded by the source corpus.

In single image deblurring, the "coarse-to-fine" scheme, i.e. gradually restoring the sharp image on different resolutions in a pyramid, is very successful in both traditional optimization-based methods and recent neural-network-based approaches. In this paper, we investigate this strategy and propose a Scale-recurrent Network (SRN-DeblurNet) for this deblurring task. Compared with the many recent learning-based approaches in [25], it has a simpler network structure, a smaller number of parameters and is easier to train. We evaluate our method on large-scale deblurring datasets with complex motion. Results show that our method can produce better quality results than state-of-the-arts, both quantitatively and qualitatively.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

ZeroWaste Dataset: Towards Deformable Object Segmentation in Cluttered Scenes (2021, methodology): “We used the SRN-Deblur [55] method to remove motion blur resulting from the fast-moving conveyor belt.”
Pre-Trained Image Processing Transformer (2020, background): “A various excellent works are also proposed for the other tasks, such as denoising [68, 32, 37, 45, 24], dehazing [6, 46, 85, 80], deraining [36, 78, 62, 29, 74, 47], and debluring [67, 53, 23, 10].”
Wide-Context Semantic Image Extrapolation (2019, background): “In regard to technical strategies, deep learning becomes popular and effective in low-level vision [8, 26, 39, 46, 48].”
Efficient Frequency Domain-based Transformers for High-Quality Image Deblurring (2022, methodology): “In recent years, we have witnessed significant advances in image deblurring due to the development of different deep CNN models [3, 4, 6, 13, 19, 28, 29].”
Spatially-Attentive Patch-Hierarchical Network for Adaptive Motion Deblurring (2020, methodology): “We provide extensive comparisons with state-of-the-art learning-based methods, namely MS-CNN[13], DeblurGAN[8], DeblurGAN-v2[9], SRN[20], and Stack(4)-DMPHN[26].”
GridDehazeNet: Attention-Based Multi-Scale Network for Image Dehazing (2019, background): “The grid network has clear advantages over the encoderdecoder network and the conventional multi-scale network extensively used in image restoration [18, 41, 38, 27].”
Uformer: A General U-Shaped Transformer for Image Restoration (2021, background): “[79], SRN [56], DMPHN [78], and MPRNet [77].”
Deblurring by Realistic Blurring (2020, methodology): “The blurry images in the ﬁrst column are from the RWBI dataset, and the images in the following columns are the re-sults of Nah et al. [21], Tao et al. [36] and the proposed DBGAN(+).”
Human-Aware Motion Deblurring (2019, methodology): “In a similar spirit, more advanced deep learning based deblurring models [35, 54, 27, 21, 58, 40] have been designed, e.”
EnlightenGAN: Deep Light Enhancement Without Paired Supervision (2019, background): “tion and enhancement approaches using deep learning heavily rely on either synthesized or captured corrupted and clean image pairs to train, such as super-resolution [6], denoising [7] and deblurring [8]. However, the availability assumption of paired training images has raised more difﬁculties, whe”
Real-World Blur Dataset for Learning and Benchmarking Deblurring Algorithms (2020, methodology): “(t) Ground truth (s) SRN-DeblurNet* [42] 29.”
Dynamic Scene Deblurring With Parameter Selective Sharing and Nested Skip Connections (2019, methodology): “Compared with [33], we make better use of parameters and design a new nested skip connection structure for the nonlinear transformation modules.”
Rethinking Coarse-to-Fine Approach in Single Image Deblurring (2021, methodology): “Motivated by the success of DeepDeblur, various CNN-based image deblurring methods [22, 23, 31, 5] have been introduced with remarkable performance improvements.”
Deep Image Deblurring: A Survey (2022, methodology): “…multi-scale architecture (Gao et al. 2019; Purohit et al. 2019; Tao et al. 2018), improving deblurring at different scales, e . g ., using nested skip connections in Gao et al. (2019), or increasing the connection of networks across different scales, e . g ., recurrent layers in Tao et al. (2018).”
Spatio-Temporal Filter Adaptive Network for Video Deblurring (2019, methodology): “More recently, many end-to-end CNN models [38, 44, 17, 23, 26] have also been proposed for image deblurring.”
Deep Stacked Hierarchical Multi-Patch Network for Image Deblurring (2019, methodology): “Different from approaches [14, 23] that evaluate the Mean Square Error (MSE) loss at each level, we evaluate the MSE loss only at the output of level 1 (which resembles res.”