Recurrent Generative Adversarial Network for Face Completion | |
Wang Q(王强)1,2,3; Fan HJ(范慧杰)1,2![]() ![]() ![]() ![]() | |
Department | 机器人学研究室 |
Source Publication | IEEE TRANSACTIONS ON MULTIMEDIA
![]() |
ISSN | 1520-9210 |
2021 | |
Volume | 23Pages:429-442 |
Indexed By | SCI |
WOS ID | WOS:000601877600034 |
Contribution Rank | 1 |
Funding Organization | National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [61873259, 61821005] ; Cooperation Projects of CAS ITRI [CAS-ITRI201905] ; Key Research and Development Program of Liaoning [2019JH2/10100014] |
Keyword | Face Feature extraction Recurrent neural networks Generative adversarial networks Semantics Image restoration Gallium nitride Recurrent neural network generative adversarial network face completion short link |
Abstract | Most recently-proposed face completion algorithms use high-level features extracted from convolutional neural networks (CNNs) to recover semantic texture content. Although the completed face is natural-looking, the synthesized content still lacks lots of high-frequency details, since the high-level features cannot supply sufficient spatial information for details recovery. To tackle this limitation, in this paper, we propose a Recurrent Generative Adversarial Network (RGAN) for face completion. Unlike previous algorithms, RGAN can take full advantage of multi-level features, and further provide advanced representations from multiple perspectives, which can well restore spatial information and details in face completion. Specifically, our RGAN model is composed of a CompletionNet and a DisctiminationNet, where the CompletionNet consists of two deep CNNs and a recurrent neural network (RNN). The first deep CNN is presented to learn the internal regulations of a masked image and represent it with multi-level features. The RNN model then exploits the relationships among the multi-level features and transfers these features in another domain, which can be used to complete the face image. Benefiting from bidirectional short links, another CNN is used to fuse multi-level features transferred from RNN and reconstruct the face image in different scales. Meanwhile, two context discrimination networks in the DisctiminationNet are adopted to ensure the completed image consistency globally and locally. Experimental results on benchmark datasets demonstrate qualitatively and quantitatively that our model performs better than the state-of-the-art face completion models, and simultaneously generates realistic image content and high-frequency details. The code will be released available soon. |
Language | 英语 |
Citation statistics | |
Document Type | 期刊论文 |
Identifier | http://ir.sia.cn/handle/173321/28143 |
Collection | 机器人学研究室 |
Corresponding Author | Fan HJ(范慧杰) |
Affiliation | 1.State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China 2.Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110016, China 3.University of Chinese Academy of Sciences, Huairou 100049, China |
Recommended Citation GB/T 7714 | Wang Q,Fan HJ,Sun G,et al. Recurrent Generative Adversarial Network for Face Completion[J]. IEEE TRANSACTIONS ON MULTIMEDIA,2021,23:429-442. |
APA | Wang Q,Fan HJ,Sun G,Ren WH,&Tang YD.(2021).Recurrent Generative Adversarial Network for Face Completion.IEEE TRANSACTIONS ON MULTIMEDIA,23,429-442. |
MLA | Wang Q,et al."Recurrent Generative Adversarial Network for Face Completion".IEEE TRANSACTIONS ON MULTIMEDIA 23(2021):429-442. |
Files in This Item: | ||||||
File Name/Size | DocType | Version | Access | License | ||
Recurrent Generative(12315KB) | 期刊论文 | 出版稿 | 开放获取 | CC BY-NC-SA | View Application Full Text |
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment