Abstract
We propose a novel framework for simultaneously generating and manipulating the face images with desired attributes. While the state-of-the-art attribute editing techniques have achieved the impressive performance for creating realistic attribute effects, they only address the image editing problem, using the input image as the condition of model. Recently, several studies attempt to tackle both novel face generation and attribute editing problem using a single model. However, their image quality is still unsatisfactory. Our goal is to develop a single unified model that can simultaneously create and edit high quality face images with desired attributes. A key idea of our work is that we decompose the image into the latent and attribute vector in low dimensional representation, and then utilize the GANs framework for mapping the low dimensional representation to the image. In this way, we can address both the generation and editing problem by training the proposed GANs, namely Editable GAN. For qualitative and quantitative evaluations, the proposed GANs outperform recent algorithms addressing the same problem. Also, we show that our model can achieve the competitive performance with the state-of-the-art attribute editing technique in terms of attribute editing quality.
| Original language | English |
|---|---|
| Title of host publication | Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers |
| Editors | Hongdong Li, C.V. Jawahar, Greg Mori, Konrad Schindler |
| Publisher | Springer Verlag |
| Pages | 39-55 |
| Number of pages | 17 |
| ISBN (Print) | 9783030208868 |
| DOIs | |
| Publication status | Published - 2019 Jan 1 |
| Event | 14th Asian Conference on Computer Vision, ACCV 2018 - Perth, Australia Duration: 2018 Dec 2 → 2018 Dec 6 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 11361 LNCS |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Conference
| Conference | 14th Asian Conference on Computer Vision, ACCV 2018 |
|---|---|
| Country | Australia |
| City | Perth |
| Period | 18/12/2 → 18/12/6 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Computer Science(all)
Cite this
}
Editable Generative Adversarial Networks : Generating and Editing Faces Simultaneously. / Baek, Kyungjune; Bang, Duhyeon; Shim, Hyunjung.
Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers. ed. / Hongdong Li; C.V. Jawahar; Greg Mori; Konrad Schindler. Springer Verlag, 2019. p. 39-55 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11361 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Editable Generative Adversarial Networks
T2 - Generating and Editing Faces Simultaneously
AU - Baek, Kyungjune
AU - Bang, Duhyeon
AU - Shim, Hyunjung
PY - 2019/1/1
Y1 - 2019/1/1
N2 - We propose a novel framework for simultaneously generating and manipulating the face images with desired attributes. While the state-of-the-art attribute editing techniques have achieved the impressive performance for creating realistic attribute effects, they only address the image editing problem, using the input image as the condition of model. Recently, several studies attempt to tackle both novel face generation and attribute editing problem using a single model. However, their image quality is still unsatisfactory. Our goal is to develop a single unified model that can simultaneously create and edit high quality face images with desired attributes. A key idea of our work is that we decompose the image into the latent and attribute vector in low dimensional representation, and then utilize the GANs framework for mapping the low dimensional representation to the image. In this way, we can address both the generation and editing problem by training the proposed GANs, namely Editable GAN. For qualitative and quantitative evaluations, the proposed GANs outperform recent algorithms addressing the same problem. Also, we show that our model can achieve the competitive performance with the state-of-the-art attribute editing technique in terms of attribute editing quality.
AB - We propose a novel framework for simultaneously generating and manipulating the face images with desired attributes. While the state-of-the-art attribute editing techniques have achieved the impressive performance for creating realistic attribute effects, they only address the image editing problem, using the input image as the condition of model. Recently, several studies attempt to tackle both novel face generation and attribute editing problem using a single model. However, their image quality is still unsatisfactory. Our goal is to develop a single unified model that can simultaneously create and edit high quality face images with desired attributes. A key idea of our work is that we decompose the image into the latent and attribute vector in low dimensional representation, and then utilize the GANs framework for mapping the low dimensional representation to the image. In this way, we can address both the generation and editing problem by training the proposed GANs, namely Editable GAN. For qualitative and quantitative evaluations, the proposed GANs outperform recent algorithms addressing the same problem. Also, we show that our model can achieve the competitive performance with the state-of-the-art attribute editing technique in terms of attribute editing quality.
UR - http://www.scopus.com/inward/record.url?scp=85066793990&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85066793990&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-20887-5_3
DO - 10.1007/978-3-030-20887-5_3
M3 - Conference contribution
AN - SCOPUS:85066793990
SN - 9783030208868
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 39
EP - 55
BT - Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers
A2 - Li, Hongdong
A2 - Jawahar, C.V.
A2 - Mori, Greg
A2 - Schindler, Konrad
PB - Springer Verlag
ER -