SCAN: Spatial and Channel Attention Normalization for Image Inpainting

Shiyu Chen, Wenxin Yu*, Liang Nie, Xuewen Zhang, Siyuan Li, Zhiqiang Zhang, Jun Gong

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

1 引用 (Scopus)

摘要

Image inpainting focuses on predicting contents with shape structure and consistent details in damaged regions. Recent approaches based on convolutional neural network (CNN) have shown promising results via adversarial learning, attention mechanism, and various loss functions. This paper introduces a novel module named Spatial and Channel Attention Normalization (SCAN), combining attention mechanisms in spatial and channel dimension and normalization to handle complex information of known regions while avoiding its misuse. Experiments on the varies datasets indicate that the performance of the proposed method outperforms the current state-of-the-art (SOTA) inpainting approaches.

源语言英语
主期刊名Neural Information Processing - 28th International Conference, ICONIP 2021, Proceedings
编辑Teddy Mantoro, Minho Lee, Media Anugerah Ayu, Kok Wai Wong, Achmad Nizar Hidayanto
出版商Springer Science and Business Media Deutschland GmbH
674-682
页数9
ISBN(印刷版)9783030923099
DOI
出版状态已出版 - 2021
活动28th International Conference on Neural Information Processing, ICONIP 2021 - Virtual, Online
期限: 8 12月 202112 12月 2021

出版系列

姓名Communications in Computer and Information Science
1517 CCIS
ISSN(印刷版)1865-0929
ISSN(电子版)1865-0937

会议

会议28th International Conference on Neural Information Processing, ICONIP 2021
Virtual, Online
时期8/12/2112/12/21

指纹

探究 'SCAN: Spatial and Channel Attention Normalization for Image Inpainting' 的科研主题。它们共同构成独一无二的指纹。

引用此