Error Compensation Framework for Flow-Guided Video Inpainting

被引:13
|
作者
Kang, Jaeyeon [1 ]
Oh, Seoung Wug [2 ]
Kim, Seon Joo [1 ]
机构
[1] Yonsei Univ, Seoul, South Korea
[2] Adobe, San Jose, CA USA
来源
关键词
Video inpainting; Object removal; Video restoration;
D O I
10.1007/978-3-031-19784-0_22
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The key to video inpainting is to use correlation information from as many reference frames as possible. Existing flow-based propagation methods split the video synthesis process into multiple steps: flow completion -> pixel propagation -> synthesis. However, there is a significant drawback that the errors in each step continue to accumulate and amplify in the next step. To this end, we propose an Error Compensation Framework for Flow-guided Video Inpainting (ECFVI), which takes advantage of the flow-based method and offsets its weaknesses. We address the weakness with the newly designed flow completion module and the error compensation network that exploits the error guidance map. Our approach greatly improves the temporal consistency and the visual quality of the completed videos. Experimental results show the superior performance of our proposed method with the speed up of x6, compared to the state-of-the-art methods. In addition, we present a new benchmark dataset for evaluation by supplementing the weaknesses of existing test datasets.
引用
下载
收藏
页码:375 / 390
页数:16
相关论文
共 50 条
  • [1] Flow-Guided Transformer for Video Inpainting
    Zhang, Kaidong
    Fu, Jingjing
    Liu, Dong
    COMPUTER VISION - ECCV 2022, PT XVIII, 2022, 13678 : 74 - 90
  • [2] Deep Flow-Guided Video Inpainting
    Xu, Rui
    Li, Xiaoxiao
    Zhou, Bolei
    Loy, Chen Change
    2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 3718 - 3727
  • [3] Local and nonlocal flow-guided video inpainting
    Jing Wang
    Zongju Yang
    Zhanqiang Huo
    Wei Chen
    Multimedia Tools and Applications, 2024, 83 : 10321 - 10340
  • [4] Towards An End-to-End Framework for Flow-Guided Video Inpainting
    Li, Zhen
    Lu, Cheng-Ze
    Qin, Jianhua
    Guo, Chun-Le
    Cheng, Ming-Ming
    2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022), 2022, : 17541 - 17550
  • [5] Local and nonlocal flow-guided video inpainting
    Wang, Jing
    Yang, Zongju
    Huo, Zhanqiang
    Chen, Wei
    MULTIMEDIA TOOLS AND APPLICATIONS, 2024, 83 (04) : 10321 - 10340
  • [6] Flow-Guided Video Inpainting with Scene Templates
    Lao, Dong
    Zhu, Peihao
    Wonka, Peter
    Sundaramoorthi, Ganesh
    2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, : 14579 - 14588
  • [7] FSTT: Flow-Guided Spatial Temporal Transformer for Deep Video Inpainting
    Liu, Ruixin
    Zhu, Yuesheng
    ELECTRONICS, 2023, 12 (21)
  • [8] FVIFormer: Flow-Guided Global-Local Aggregation Transformer Network for Video Inpainting
    Yan, Weiqing
    Sun, Yiqiu
    Yue, Guanghui
    Zhou, Wei
    Liu, Hantao
    IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS, 2024, 14 (02) : 235 - 244
  • [9] Flow-Guided Single Object Tracking Framework In UAV Aerial Video
    Zhu, Wenjun
    Yu, Xi
    Meng, Jun
    PROCEEDINGS OF THE 32ND 2020 CHINESE CONTROL AND DECISION CONFERENCE (CCDC 2020), 2020, : 461 - 468
  • [10] Flow-Guided Transformer for Video Colorization
    Zhai, Yan
    Tao, Zhulin
    Dai, Longquan
    Wang, He
    Huang, Xianglin
    Yang, Lifang
    Proceedings - International Conference on Image Processing, ICIP, 2023, : 2485 - 2489