Research Article
A Survey of Face Image Inpainting Based on Deep Learning
@INPROCEEDINGS{10.1007/978-3-030-99191-3_7, author={Shiqi Su and Miao Yang and Libo He and Xiaofeng Shao and Yuxuan Zuo and Zhenping Qiang}, title={A Survey of Face Image Inpainting Based on Deep Learning}, proceedings={Cloud Computing. 11th EAI International Conference, CloudComp 2021, Virtual Event, December 9--10, 2021, Proceedings}, proceedings_a={CLOUDCOMP}, year={2022}, month={3}, keywords={Face inpainting Deep learning Attention inpainting Semantic inpainting}, doi={10.1007/978-3-030-99191-3_7} }- Shiqi Su
Miao Yang
Libo He
Xiaofeng Shao
Yuxuan Zuo
Zhenping Qiang
Year: 2022
A Survey of Face Image Inpainting Based on Deep Learning
CLOUDCOMP
Springer
DOI: 10.1007/978-3-030-99191-3_7
Abstract
In recent years, deep learning has become the mainstream method of image inpainting. It can not only repair the texture of the image, obtain high-level abstract features of the image, but also recover semantic images such as human faces. Among these methods, attention mechanisms, semantic methods, and progressive networks have become very promising image inpainting models. These models implement end-to-end image inpainting and generate visually reasonable and clear image structure and texture. This paper briefly describes the face inpainting technology and summarizes the existing face image inpainting methods. We try to collect most of the face inpainting methods based on deep learning, divide them into attentional, semantic-based, and progressive inpainting networks, and prorate the methods proposed by researchers in each category in recent years. Then we summarize the dataset proposed by the predecessors and the evaluation index of the algorithm performance. Finally, we summarize the current situation and future development trends of face inpainting.
