Dual Residual Networks Leveraging the Potential of Paired Operations for Image Restoration

By Xing Liu, Masanori Suganuma, Zhun Sun, Takayuki Okatani. 2019

266 citations recorded by the source corpus.

In this paper, we study design of deep neural networks for tasks of image restoration. We propose a novel style of residual connections dubbed "dual residual connection", which exploits the potential of paired operations, e.g., up- and down-sampling or convolution with large- and small-size kernels. We design a modular block implementing this connection style; it is equipped with two containers to which arbitrary paired operations are inserted. Adopting the "unraveled" view of the residual networks proposed by Veit et al., we point out that a stack of the proposed modular blocks allows the first operation in a block interact with the second operation in any subsequent blocks. Specifying the two operations in each of the stacked blocks, we build a complete network for each individual task of image restoration. We experimentally evaluate the proposed approach on five image restoration tasks using nine datasets. The results show that the proposed networks with properly chosen paired operations outperform previous methods on almost all of the tasks and datasets.

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

Memory-Efficient Hierarchical Neural Architecture Search for Image Denoising (2019, methodology): “DuRNP [22] proposes a new style of residual connection, where two residual connections are employed to exploit the potential of paired operations.”
MAXIM: Multi-Axis MLP for Image Processing (2022, methodology): “Visual comparisons for raindrop removal on Raindrop-A [71] among AGAN [71], DuRN [55], Quan [73], and MAXIM-2S.”
Multi-Scale Boosted Dehazing Network With Dense Feature Fusion (2020, methodology): “Methods DCP [3] NLD [1] AODNet [6] MSCNN [13] MsPPN [19] DcGAN [8] GFN [14] GCANet [2] PFFNet [12] GDN [10] DuRN [11] Ours”
CLEARER: Multi-Scale Neural Architecture Search for Image Restoration (2020, methodology): “In details, BM3D [4], WNNM [8], RED [27], MemNet [40], NLRN [19], DuRN-P [24], N3Net [31], HiNAS [46], and E-CAE [38].”
Restormer: Efficient Transformer for High-Resolution Image Restoration (2021, background): “Similarly, skip connection based approaches have been shown to be effective for restoration due to specific focus on learning residual signals [24, 48, 92, 106].”
Deep Learning on Image Denoising: An overview (2019, methodology): “Other effective denoising methods are shown in [140, 105, 66].”
Restoring Vision in Adverse Weather Conditions With Patch-Based Denoising Diffusion Models (2022, methodology): “We perform comparisons of our weather-specific models with several state-of-the-art methods discussed in Section II-B for image desnowing [3], [43], [44], [58], [59], combined image deraining and dehazing [14], [45], [46], [53], [54], and removing raindrops [6], [12], [45], [55], [56].”
Uformer: A General U-Shaped Transformer for Image Restoration (2021, methodology): “For example, ResNet-based structure has been widely used for general image restoration [39, 80] as well as for speciﬁc tasks in image restoration such as super-resolution [36,81] and image denoising [15,78].”