中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 模式分析与学习团队
Image Inpainting With Local and Global Refinement

文献类型：期刊论文

作者Quan, Weize4,5; Zhang, Ruisong4,5; Zhang, Yong3; Li, Zhifeng1; Wang, Jue2; Yan, Dong-Ming4,5
刊名IEEE TRANSACTIONS ON IMAGE PROCESSING
出版日期2022
卷号31页码:2405-2420
ISSN号1057-7149
关键词Image reconstruction Generators Convolution Semantics Image edge detection Deep learning Task analysis Image inpainting neural networks receptive field
DOI10.1109/TIP.2022.3152624
通讯作者Yan, Dong-Ming(yandongming@gmail.com)
英文摘要Image inpainting has made remarkable progress with recent advances in deep learning. Popular networks mainly follow an encoder-decoder architecture (sometimes with skip connections) and possess sufficiently large receptive field, i.e., larger than the image resolution. The receptive field refers to the set of input pixels that are path-connected to a neuron. For image inpainting task, however, the size of surrounding areas needed to repair different kinds of missing regions are different, and the very large receptive field is not always optimal, especially for the local structures and textures. In addition, a large receptive field tends to involve more undesired completion results, which will disturb the inpainting process. Based on these insights, we rethink the process of image inpainting from a different perspective of receptive field, and propose a novel three-stage inpainting framework with local and global refinement. Specifically, we first utilize an encoder-decoder network with skip connection to achieve coarse initial results. Then, we introduce a shallow deep model with small receptive field to conduct the local refinement, which can also weaken the influence of distant undesired completion results. Finally, we propose an attention-based encoder-decoder network with large receptive field to conduct the global refinement. Experimental results demonstrate that our method outperforms the state of the arts on three popular publicly available datasets for image inpainting. Our local and global refinement network can be directly inserted into the end of any existing networks to further improve their inpainting performance. Code is available at https://github.com/weizequan/LGNet.git.
资助项目National Natural Science Foundation of China[62102418] ; National Natural Science Foundation of China[62172415] ; National Key Research and Development Program of China[2019YFB2204104] ; Tencent AI Laboratory Rhino-Bird Focused Research Program[JR202127] ; Open Project Program of National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University[2021SCUVS002]
WOS研究方向Computer Science ; Engineering
语种英语
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号WOS:000769973200007
资助机构National Natural Science Foundation of China ; National Key Research and Development Program of China ; Tencent AI Laboratory Rhino-Bird Focused Research Program ; Open Project Program of National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University
源URL[http://ir.ia.ac.cn/handle/173211/48123]  
专题自动化研究所_模式识别国家重点实验室_模式分析与学习团队
模式识别国家重点实验室_三维可视计算
通讯作者Yan, Dong-Ming
作者单位1.Tencent Data Platform, Shenzhen 518054, Peoples R China
2.Tencent AI Lab, Visual Comp Ctr, Shenzhen 518054, Peoples R China
3.Tencent AI Lab, Shenzhen 518054, Peoples R China
4.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
5.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
推荐引用方式
GB/T 7714		 Quan, Weize,Zhang, Ruisong,Zhang, Yong,et al. Image Inpainting With Local and Global Refinement[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2022,31:2405-2420.
APA Quan, Weize,Zhang, Ruisong,Zhang, Yong,Li, Zhifeng,Wang, Jue,&Yan, Dong-Ming.(2022).Image Inpainting With Local and Global Refinement.IEEE TRANSACTIONS ON IMAGE PROCESSING,31,2405-2420.
MLA Quan, Weize,et al."Image Inpainting With Local and Global Refinement".IEEE TRANSACTIONS ON IMAGE PROCESSING 31(2022):2405-2420.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。