Image Super-Resolution via Deep Recursive Residual Network

By Ying Tai, Jian Yang, Xiaoming Liu. 2017

2,201 citations recorded by the source corpus.

Recently, Convolutional Neural Network (CNN) based models have achieved great success in Single Image Super-Resolution (SISR). Owing to the strength of deep networks, these CNN models learn an effective nonlinear mapping from the low-resolution input image to the high-resolution target image, at the cost of requiring enormous parameters. This paper proposes a very deep CNN model (up to 52 convolutional layers) named Deep Recursive Residual Network (DRRN) that strives for deep yet concise networks. Specifically, residual learning is adopted, both in global and local manners, to mitigate the difficulty of training very deep networks, recursive learning is used to control the model parameters while increasing the depth. Extensive benchmark evaluation shows that DRRN significantly outperforms state of the art in SISR, while utilizing far fewer parameters. Code is available at https://github.com/tyshiwo/DRRN_CVPR17.

Subfields: Computer Vision and Pattern Recognition, Media Technology, Generative Models

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

In Media Technology

In Generative Models

Citations in the corpus

Image Super-Resolution via Iterative Refinement (2021, neutral): “Numerous super-resolution methods have been proposed [16], [30], [31], [32], [49], [50], [51], [52], [53].”
Hierarchical Neural Architecture Search for Single Image Super-Resolution (2020, methodology): “In this experiment, we compare our method with several hand-crafted SR models, including the Bicubic, SRCNN [7], VDSR [27], DRCN [28], DRRN [29], SelNet [30], CARN [31].”
Fast, Accurate and Lightweight Super-Resolution with Neural Architecture Search (2019, methodology): “Moreover, it achieves higher PSNR and SSIM with fewer FLOPS than VDSR [ Kim et al. , 2016a], DRRN [Tai et al. , 2017a] and many others.”
Transformers in Vision: A Survey (2021, methodology): “While the SR methods [75, 127, 188, 247, 248] that are based on pixel-wise loss functions (e.”
Deep Image Prior (2017, methodology): “Qualitative comparison with bicubic upsampling and state-of-the art learning-based methods SRResNet [19], LapSRN [29] is presented in fig.”
ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (2018, background): “orks. Later on, the eld has witnessed a variety of network architectures, such as a deeper network with residual learning [5], Laplacian pyramid structure [6], residual blocks [1], recursive learning [7,8], densely connected network [9], deep back projection [10] and residual dense network [11]. Spe”
Checkmate: Breaking the Memory Wall with Optimal Tensor Rematerialization (2019, background): “Deep learning is rapidly pushing the limits of memory capacity on neural network accelerators as researchers train neural networks on high-resolution images [15, 32, 55], 3D point-clouds [10, 59], and long Natural Language Processing (NLP) sequence data [11, 14, 56].”
Image Super-Resolution Using Very Deep Residual Channel Attention Networks (2018, background): “HR Bicubic SRCNN [1] FSRCNN [2] SCN [3] VDSR [4] DRRN [5]”
Multi-Objective Reinforced Evolution in Mobile Neural Architecture Search (2019, background): “Besides, both MoreMNAS-B and D perform very competitively against DRRN (Tai et al., 2017) with much fewer multi-adds on Urban100 test set, which is a very challenging task involving rich details.”
Attention-guided CNN for image denoising (2020, methodology): “Further, the fact above is verified by Table 2, where ‘ADNet’ achieves greater PSNR value [43] than that of ‘FEB+AB’.”
Multilevel and Multiscale Network for Single-Image Super-Resolution (2019, methodology): “To reduce the network parameters, the DRRN [14] was proposed.”
Learned Image Downscaling for Upscaling Using Content Adaptive Resampler (2019, methodology): “SR models in the ‘Bicubic’ group include the SRCNN [28], VDSR [29], DRRN [55], MemNet [56], DnCNN [57], LapSRN [38], ZSSR [58], CARN [59], SRRAM [60].”
Progressive Image Deraining Networks: A Better and Simpler Baseline (2019, methodology): “In comparison, we take advantage of inter-stage recursive computation [15, 20, 28] by requiring multiple stages share the same network parameters.”
Lightweight Image Super-Resolution with Adaptive Weighted Learning Network (2019, methodology): “DRRN [11] shared parameters through recursive mechanism to not only reduce the parameters, but also improve the reconstruction quality of SR images.”
DARTS-: Robustly Stepping out of Performance Collapse Without Indicators (2020, background): “Multiple works (Ren et al., 2015; Wei et al., 2017; Tai et al., 2017; Li et al., 2018b) also empirically demonstrate that deep residual network can achieve better performance on various tasks.”
Face De-Spoofing: Anti-Spoofing via Noise Modeling (2018, methodology): “For de-mosaicing and super-resolution, many models, such as ResNet in [14,15] and joint models in [16,17,35], are learnt from the given pairs of low-quality input and high-quality ground truth.”
Efficient Long-Range Attention Network for Image Super-resolution (2022, methodology): “The ELAN can be optimized with the commonly used loss functions for SR, such as L 2 [11, 20, 48, 49], L 1 [22, 29, 62] and perceptual losses [16, 45].”
Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks (2017, methodology): “The DRRN approach [13] further trains a 52-layer network by extending the local residual learning approach of the ResNet [6] with”
Deep Learning for Image Super-Resolution: A Survey (2019, methodology): “[77] feature similarity MS-SSIM [74] multi-scale SSIM IFC [75] information ﬁdelity criterion PSNR peak signal-to-noise ratio IQA image quality assessment SSIM [62] structural similarity MOS mean opinion score VIF [76] visual information ﬁdelity MSSIM [62] mean SSIM BN [140] batch normalization GAN [”