SCGA-Net: Skip Connections Global Attention Network for Image Restoration

dc.contributor.authorRen, Dongdongen_US
dc.contributor.authorLi, Jinbaoen_US
dc.contributor.authorHan, Mengen_US
dc.contributor.authorShu, Mingleien_US
dc.contributor.editorEisemann, Elmar and Jacobson, Alec and Zhang, Fang-Lueen_US
dc.date.accessioned2020-10-29T18:51:09Z
dc.date.available2020-10-29T18:51:09Z
dc.date.issued2020
dc.description.abstractDeep convolutional neural networks (DCNN) have shown their advantages in the image restoration tasks. But most existing DCNN-based methods still suffer from the residual corruptions and coarse textures. In this paper, we propose a general framework ''Skip Connections Global Attention Network'' to focus on the semantics delivery from shallow layers to deep layers for low-level vision tasks including image dehazing, image denoising, and low-light image enhancement. First of all, by applying dense dilated convolution and multi-scale feature fusion mechanism, we establish a novel encoder-decoder network framework to aggregate large-scale spatial context and enhance feature reuse. Secondly, the solution we proposed for skipping connection uses attention mechanism to constraint information, thereby enhancing the high-frequency details of feature maps and suppressing the output of corruptions. Finally, we also present a novel attention module dubbed global constraint attention, which could effectively captures the relationship between pixels on the entire feature maps, to obtain the subtle differences among pixels and produce an overall optimal 3D attention maps. Extensive experiments demonstrate that the proposed method achieves significant improvements over the state-of-the-art methods in image dehazing, image denoising, and low-light image enhancement.en_US
dc.description.number7
dc.description.sectionheadersImage Restoration
dc.description.seriesinformationComputer Graphics Forum
dc.description.volume39
dc.identifier.doi10.1111/cgf.14163
dc.identifier.issn1467-8659
dc.identifier.pages507-518
dc.identifier.urihttps://doi.org/10.1111/cgf.14163
dc.identifier.urihttps://diglib.eg.org:443/handle/10.1111/cgf14163
dc.publisherThe Eurographics Association and John Wiley & Sons Ltd.en_US
dc.subjectComputing methodologies
dc.subjectImage processing
dc.titleSCGA-Net: Skip Connections Global Attention Network for Image Restorationen_US
Files
Collections