Multi-scale attention network for image inpainting

Qin, Jia; Bai, Huihui; Zhao, Yao

doi:10.1016/j.cviu.2020.103155

Cited by 26 publications

(6 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep learning has made significant breakthroughs in a target recognition, target classification, mural segmentation, and target tracking [6]. For example, Qin et al [7] proposed a restoration model based on multiscale attention networks to improve the authenticity of restored images by introducing multiscale attention groups. Zeng et al [8] proposed a restoration network based on a context encoder to complete the restoration of broken images by encoding the contextual semantics of full-resolution inputs.…”

Section: Related Studiesmentioning

confidence: 99%

Digital inpainting of mural images based on DC-CycleGAN

Xu,

Zhang,

2023

Herit Sci

View full text Add to dashboard Cite

Located in Dunhuang, northwest China, the Mogao Grottoes are a cultural treasure of China and the world. However, after more than 2000 years of weathering and destruction, many murals faded and were damaged. This treasure of human art is in danger. Mural inpainting through deep learning can permanently preserve mural information. Therefore, a digital restoration method combining the Deformable Convolution (DCN), ECANet, ResNet and Cycle Generative Adversarial Network (CycleGAN) is proposed. We name it DC-CycleGAN. Compared with other image digital inpainting methods, the proposed DC-CycleGAN based mural image color inpainting method has better inpainting effects and higher model performance. Compared with the current repair network, the Frechet Inception Distance (FID) value and the two-image structural similarity metric (SSIM) value are increased by 52.61% and 7.08%, respectively. Image color inpainting of Dunhuang murals can not only protect and inherit Chinese culture, but also promote academic research and development in related fields.

show abstract

Section: Related Studiesmentioning

confidence: 99%

Digital inpainting of mural images based on DC-CycleGAN

Xu,

Zhang,

2023

Herit Sci

View full text Add to dashboard Cite

show abstract

“…The first is conceptualization. That is, the concepts related to certain objective phenomena are abstracted into models, but the model's performance has nothing to do with the specific environment (Won & Sim, 2020;Chen et al, 2021;Qin et al, 2021). The second is clearness.…”

Section: Ontology-based Supply Chain Scenario Modelingmentioning

confidence: 99%

Internet of Things-Enabled Tourism Economic Data Analysis and Supply Chain Modeling

Wang

Fang

2022

Technological and Economic Development of Economy

View full text Add to dashboard Cite

The purpose is to cut the costs of Supply Chain enterprises in Ice-Snow Tourism (IST) and improve the intelligence and automation of Supply Chain Management (SCM). First, the spatial-temporal characteristics of economic data of the IST Supply Chain are analyzed based on the Internet of Things (IoT). Second, the annual Online Public Attention (OPA) data to IST in domestic cities and regions are collected. The quarterly concentration index and Gini coefficient are used to analyze their spatial and temporal characteristics. Then, the weighted fusion algorithm used for the Supply Chain scenario modeling is improved to solve data redundancy and improve information accuracy. Finally, the framework of the IST-oriented Supply Chain scenario ontology model is proposed. The experimental results show that Internet users give much attention to IST from 2011 to 2021. OPA to IST increased first and decreased and peaked in 2016. The final fusion value of the proposed data fusion algorithm is 20.0221, and that of the adaptive Weighted Average Method (WAM) is 20.0724. Thus, the proposed algorithm outperforms the adaptive WAM. The traditional scenario-based ontology model takes people as the center. In contrast, the Supply Chain scenario-based ontology model centers around product state and scenario. Therefore, the proposed Supply Chain scenario-based ontology model is entirely new. The proposed scenariobased ontology model using polymorphic IoT lays the foundation for developing an intelligent and automatic SCM. It has great practical significance in realizing efficient tourism industry management and SCM.

show abstract

“…Liu et al [39] used multiple residual subnets with different receptive fields to extract multi-scale features, and proposed a dual-branch module to fuse inter and inner correlations of multi-scale features. Some researchers [22], [40], [41] proposed multi-column subnets and multi-scale attentions to generate more realistic and complex results. Zeng et al [42] learned pyramid-context encoder and multi-scale decoder to restore the image.…”

Section: B Learning-based Image Inpaintingmentioning

confidence: 99%

Multi-Reception and Multi-Gradient Discriminator for Image Inpainting

et al. 2022

View full text Add to dashboard Cite

Many deep learning methods for image inpainting rely on the encoder-decoder architecture to estimate missing contents. When guidance information from uncorrupted regions could not be adequately represented or utilized, the encoder may have difficulty handling the rich surrounding or background pixels, and the decoder could not recover visually sophisticated or realistic contents. This paper proposes an effective multi-scale optimization network to alleviate these issues and generate coherent results with fine details. It encodes multi-receptive fields feature maps adaptively and puts multi-scale outputs into a discriminator to guide training. Specifically, we propose a Multi-Receptive feature maps & masks Selective Fusion (MRSF) operator that can adaptively extract features at different receptive fields to handle sophisticated destroyed images. Then a Multi-Gradients Discriminator (MGD) module uses intermediate features of the discriminator to guide the generator to produce results with natural textures and semantically real contents. Experiments on several benchmark datasets demonstrate that our proposed method can synthesize more realistic and coherent image contents.

show abstract

Multi-scale attention network for image inpainting

Cited by 26 publications

References 36 publications

Digital inpainting of mural images based on DC-CycleGAN

Digital inpainting of mural images based on DC-CycleGAN

Internet of Things-Enabled Tourism Economic Data Analysis and Supply Chain Modeling

Multi-Reception and Multi-Gradient Discriminator for Image Inpainting

Contact Info

Product

Resources

About