Progressive Image Deraining Networks: A Better and Simpler Baseline

By Dongwei Ren, Wangmeng Zuo, Qinghua Hu, Peng Fei Zhu, Deyu Meng. Computer Vision and Pattern Recognition, 2019

1,016 citations recorded by the source corpus.

Along with the deraining performance improvement of deep networks, their structures and learning become more and more complicated and diverse, making it difficult to analyze the contribution of various network modules when developing new deraining networks. To handle this issue, this paper provides a better and simpler baseline deraining network by considering network architecture, input and output, and loss functions. Specifically, by repeatedly unfolding a shallow ResNet, progressive ResNet (PRN) is proposed to take advantage of recursive computation. A recurrent layer is further introduced to exploit the dependencies of deep features across stages, forming our progressive recurrent network (PReNet). Furthermore, intra-stage recursive computation of ResNet can be adopted in PRN and PReNet to notably reduce network parameters with unsubstantial degradation in deraining performance. For network input and output, we take both stage-wise result and original rainy image as input to each ResNet and finally output the prediction of residual image. As for loss functions, single MSE or negative SSIM losses are sufficient to train PRN and PReNet. Experiments show that PRN and PReNet perform favorably on both synthetic and real rainy images. Considering its simplicity, efficiency and effectiveness, our models are expected to serve as a suitable baseline in future deraining research. The source codes are available at https://github.com/csdwren/PReNet.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/6d411f15a0d17aba58006aa7d2eaebc66a0ca8c8

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

A Model-Driven Deep Neural Network for Single Image Rain Removal (2020, methodology): “[35] designed a PReNet that repeatedly unfolded several Resblocks and a LSTM layer.”
HINet: Half Instance Normalization Network for Image Restoration (2021, background): “[35] proposes a progressive recurrent network by repeatedly unfolding a shallow ResNet [15], and introduces a recurrent layer to exploit the dependencies of deep features across stages.”
Attention-guided CNN for image denoising (2020, neutral): “object detection [49] and image restoration [5, 40].”
Image Restoration with Mean-Reverting Stochastic Differential Equations (2023, methodology): “Moreover, we compare our methods with several state-of-the-art deraining approaches such as JORDER (Yang et al., 2019), PReNet (Ren et al., 2019), MPRNet (Zamir et al., 2021), and MAXIM (Tu et al., 2022).”
CycleISP: Real Image Restoration via Improved Data Synthesis (2020, methodology): “Motivated by the advances of recent low-level vision methods [48, 63, 64, 66] based on the residual learning framework [29], we propose the RRG module, as shown in Fig.”
Deep Generalized Unfolding Networks for Image Restoration (2022, methodology): “We compare our proposed DGUNet with eight competitive methods [19,31,38,54,68,72,79,81].”
Image De-raining via Continual Learning (2021, methodology): “PreNet [29], PRN [29], NLEDN [22], REHEN [42] as well as SASI [36].”
Image De-Raining Transformer (2022, methodology): “Ren et al. established a simple yet efﬁcient network by unfolding a shallow ResNet with intra-stage recursive computation [72].”
Multi-Scale Progressive Fusion Network for Single Image Deraining (2020, methodology): “…select a total of 850 samples from COCO [3] and BDD [38] datasets to create three new synthetic rain datasets COCO350 (for detection), BDD350 (for performing deraining methods [19, 26], the restoration procedures are directly implemented on these three datasets to produce the rain-free images.”
Restormer: Efficient Transformer for High-Resolution Image Restoration (2021, methodology): “Rainy Image RESCAN [43] PreNet [64] MSPFN [32] MPRNet [93] Restormer”
Uformer: A General U-Shaped Transformer for Image Restoration (2021, methodology): “We conduct the deraining experiments on SPAD [14] and compare with 7 deraining methods: GMM [59], DDN [60], RESCAN [31], PReNet [61], JORDER-E [62], SPANet [14], and RCDNet [15].”
Deep Learning on Image Denoising: An overview (2019, methodology): “Changing network architectures involves the following methods [246, 149, 177]: fusing features from multiple inputs of a CNN; changing the loss function; increasing depth or width of the CNN; adding some auxiliary plug-ins into CNNs; and introducing skip connections or cascade operations into CNNs.”
Multi-Stage Progressive Image Restoration (2021, methodology): “Rainy Image UMRL [82] RESCAN [47] PreNet [63] MSPFN [37] MPRNet (Ours)”
MAXIM: Multi-Axis MLP for Image Processing (2022, methodology): “Input Target RESCAN [48] PreNet [75] MSPFN [33] MPRNet [108] HINet [15] MAXIM-2S”