Face Inpainting via Nested Generative Adversarial Networks
Face inpainting aims to repaired damaged images caused by occlusion or cover. In recent years, deep learning based approaches have shown promising results for the challenging task of image inpainting. However, there are still limitation in reconstructing reasonable structures because of over-smoothe...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8883161/ |
| _version_ | 1818444490933796864 |
|---|---|
| author | Zhijiang Li Haonan Zhu Liqin Cao Lei Jiao Yanfei Zhong Ailong Ma |
| author_facet | Zhijiang Li Haonan Zhu Liqin Cao Lei Jiao Yanfei Zhong Ailong Ma |
| author_sort | Zhijiang Li |
| collection | DOAJ |
| description | Face inpainting aims to repaired damaged images caused by occlusion or cover. In recent years, deep learning based approaches have shown promising results for the challenging task of image inpainting. However, there are still limitation in reconstructing reasonable structures because of over-smoothed and/or blurred results. The distorted structures or blurred textures are inconsistent with surrounding areas and require further post-processing to blend the results. In this paper, we present a novel generative model-based approach, which consisted by nested two Generative Adversarial Networks (GAN), the sub-confrontation GAN in generator and parent-confrontation GAN. The sub-confrontation GAN, which is in the image generator of parent-confrontation GAN, can find the location of missing area and reduce mode collapse as a prior constraint. To avoid generating vague details, a novel residual structure is designed in the sub-confrontation GAN to deliver richer original image information to the deeper layers. The parent-confrontation GAN includes an image generation part and a discrimination part. The discrimination part of parent-confrontation GAN includes global and local discriminator, which benefits the reconstruction of overall coherency of the repaired image while obtaining local details. The experiments are executed over the publicly available dataset CelebA, and the results show that our method outperforms current state-of-the-art techniques quantitatively and qualitatively. |
| first_indexed | 2024-12-14T19:16:46Z |
| format | Article |
| id | doaj.art-3c0cfb11592147b7b9356f517b32a3bf |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-14T19:16:46Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-3c0cfb11592147b7b9356f517b32a3bf2022-12-21T22:50:34ZengIEEEIEEE Access2169-35362019-01-01715546215547110.1109/ACCESS.2019.29496148883161Face Inpainting via Nested Generative Adversarial NetworksZhijiang Li0Haonan Zhu1Liqin Cao2https://orcid.org/0000-0003-2945-2708Lei Jiao3https://orcid.org/0000-0002-7115-6489Yanfei Zhong4https://orcid.org/0000-0001-9446-5850Ailong Ma5School of Printing and Packaging, Wuhan University, Wuhan, ChinaSchool of Printing and Packaging, Wuhan University, Wuhan, ChinaSchool of Printing and Packaging, Wuhan University, Wuhan, ChinaDepartment of Information and Communication Technology, University of Agder, Grimstad, NorwayState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, ChinaFace inpainting aims to repaired damaged images caused by occlusion or cover. In recent years, deep learning based approaches have shown promising results for the challenging task of image inpainting. However, there are still limitation in reconstructing reasonable structures because of over-smoothed and/or blurred results. The distorted structures or blurred textures are inconsistent with surrounding areas and require further post-processing to blend the results. In this paper, we present a novel generative model-based approach, which consisted by nested two Generative Adversarial Networks (GAN), the sub-confrontation GAN in generator and parent-confrontation GAN. The sub-confrontation GAN, which is in the image generator of parent-confrontation GAN, can find the location of missing area and reduce mode collapse as a prior constraint. To avoid generating vague details, a novel residual structure is designed in the sub-confrontation GAN to deliver richer original image information to the deeper layers. The parent-confrontation GAN includes an image generation part and a discrimination part. The discrimination part of parent-confrontation GAN includes global and local discriminator, which benefits the reconstruction of overall coherency of the repaired image while obtaining local details. The experiments are executed over the publicly available dataset CelebA, and the results show that our method outperforms current state-of-the-art techniques quantitatively and qualitatively.https://ieeexplore.ieee.org/document/8883161/Face inpaintingdeep neural networknested GAN |
| spellingShingle | Zhijiang Li Haonan Zhu Liqin Cao Lei Jiao Yanfei Zhong Ailong Ma Face Inpainting via Nested Generative Adversarial Networks IEEE Access Face inpainting deep neural network nested GAN |
| title | Face Inpainting via Nested Generative Adversarial Networks |
| title_full | Face Inpainting via Nested Generative Adversarial Networks |
| title_fullStr | Face Inpainting via Nested Generative Adversarial Networks |
| title_full_unstemmed | Face Inpainting via Nested Generative Adversarial Networks |
| title_short | Face Inpainting via Nested Generative Adversarial Networks |
| title_sort | face inpainting via nested generative adversarial networks |
| topic | Face inpainting deep neural network nested GAN |
| url | https://ieeexplore.ieee.org/document/8883161/ |
| work_keys_str_mv | AT zhijiangli faceinpaintingvianestedgenerativeadversarialnetworks AT haonanzhu faceinpaintingvianestedgenerativeadversarialnetworks AT liqincao faceinpaintingvianestedgenerativeadversarialnetworks AT leijiao faceinpaintingvianestedgenerativeadversarialnetworks AT yanfeizhong faceinpaintingvianestedgenerativeadversarialnetworks AT ailongma faceinpaintingvianestedgenerativeadversarialnetworks |