Discovering interpretable latent space directions of gans beyond binary attributes

Generative adversarial networks (GANs) learn to map noise latent vectors to high- fidelity image outputs. It is found that the input latent space shows semantic correlations with the output image space. Recent works aim to interpret the latent space and discover meaningful directions that correspond...

Full description

Saved in:
Bibliographic Details
Main Authors: YANG, Huiting, CHAI, Liangyu, WEN, Qiang, ZHAO, Shuang, SUN, Zixun, HE, Shengfeng
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2021
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/8521
https://ink.library.smu.edu.sg/context/sis_research/article/9524/viewcontent/Discovering_Interpretable_Latent_Space_Directions_of_GANs_Beyond_Binary_Attributes.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
id sg-smu-ink.sis_research-9524
record_format dspace
spelling sg-smu-ink.sis_research-95242024-01-22T15:02:22Z Discovering interpretable latent space directions of gans beyond binary attributes YANG, Huiting CHAI, Liangyu WEN, Qiang ZHAO, Shuang SUN, Zixun HE, Shengfeng Generative adversarial networks (GANs) learn to map noise latent vectors to high- fidelity image outputs. It is found that the input latent space shows semantic correlations with the output image space. Recent works aim to interpret the latent space and discover meaningful directions that correspond to human interpretable image transformations. However, these methods either rely on explicit scores of attributes (e.g., memorability) or are restricted to binary ones (e.g., gender), which largely limits the applicability of editing tasks, especially for free- form artistic tasks like style/anime editing. In this paper, we propose an adversarial method, AdvStyle, for discovering interpretable directions in the absence of well- labeled scores or binary attributes. In particular, the proposed adversarial method simultaneously optimizes the discovered directions and the attribute assessor using the target attribute data as positive samples, while the generated ones being negative. In this way, arbitrary attributes can be edited by collecting positive data only, and the proposed method learns a controllable representation enabling manipulation of non- binary attributes like anime styles and facial characteristics. Moreover, the proposed learning strategy attenuates the entanglement between attributes, such that multi-attribute manipulation can be easily achieved without any additional constraint. Furthermore, we reveal several interesting semantics with the involuntarily learned negative directions. Extensive experiments on 9 anime attributes and 7 human attributes demonstrate the effectiveness of our adversarial approach qualitatively and quantitatively 2021-06-01T07:00:00Z text application/pdf https://ink.library.smu.edu.sg/sis_research/8521 info:doi/10.1109/CVPR46437.2021.01200 https://ink.library.smu.edu.sg/context/sis_research/article/9524/viewcontent/Discovering_Interpretable_Latent_Space_Directions_of_GANs_Beyond_Binary_Attributes.pdf http://creativecommons.org/licenses/by-nc-nd/4.0/ Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Attribute data Binary attributes Freeforms High-fidelity images Image space Image transformations Latent vectors Learn+ Positive data Space directions Databases and Information Systems
institution Singapore Management University
building SMU Libraries
continent Asia
country Singapore
Singapore
content_provider SMU Libraries
collection InK@SMU
language English
topic Attribute data
Binary attributes
Freeforms
High-fidelity images
Image space
Image transformations
Latent vectors
Learn+
Positive data
Space directions
Databases and Information Systems
spellingShingle Attribute data
Binary attributes
Freeforms
High-fidelity images
Image space
Image transformations
Latent vectors
Learn+
Positive data
Space directions
Databases and Information Systems
YANG, Huiting
CHAI, Liangyu
WEN, Qiang
ZHAO, Shuang
SUN, Zixun
HE, Shengfeng
Discovering interpretable latent space directions of gans beyond binary attributes
description Generative adversarial networks (GANs) learn to map noise latent vectors to high- fidelity image outputs. It is found that the input latent space shows semantic correlations with the output image space. Recent works aim to interpret the latent space and discover meaningful directions that correspond to human interpretable image transformations. However, these methods either rely on explicit scores of attributes (e.g., memorability) or are restricted to binary ones (e.g., gender), which largely limits the applicability of editing tasks, especially for free- form artistic tasks like style/anime editing. In this paper, we propose an adversarial method, AdvStyle, for discovering interpretable directions in the absence of well- labeled scores or binary attributes. In particular, the proposed adversarial method simultaneously optimizes the discovered directions and the attribute assessor using the target attribute data as positive samples, while the generated ones being negative. In this way, arbitrary attributes can be edited by collecting positive data only, and the proposed method learns a controllable representation enabling manipulation of non- binary attributes like anime styles and facial characteristics. Moreover, the proposed learning strategy attenuates the entanglement between attributes, such that multi-attribute manipulation can be easily achieved without any additional constraint. Furthermore, we reveal several interesting semantics with the involuntarily learned negative directions. Extensive experiments on 9 anime attributes and 7 human attributes demonstrate the effectiveness of our adversarial approach qualitatively and quantitatively
format text
author YANG, Huiting
CHAI, Liangyu
WEN, Qiang
ZHAO, Shuang
SUN, Zixun
HE, Shengfeng
author_facet YANG, Huiting
CHAI, Liangyu
WEN, Qiang
ZHAO, Shuang
SUN, Zixun
HE, Shengfeng
author_sort YANG, Huiting
title Discovering interpretable latent space directions of gans beyond binary attributes
title_short Discovering interpretable latent space directions of gans beyond binary attributes
title_full Discovering interpretable latent space directions of gans beyond binary attributes
title_fullStr Discovering interpretable latent space directions of gans beyond binary attributes
title_full_unstemmed Discovering interpretable latent space directions of gans beyond binary attributes
title_sort discovering interpretable latent space directions of gans beyond binary attributes
publisher Institutional Knowledge at Singapore Management University
publishDate 2021
url https://ink.library.smu.edu.sg/sis_research/8521
https://ink.library.smu.edu.sg/context/sis_research/article/9524/viewcontent/Discovering_Interpretable_Latent_Space_Directions_of_GANs_Beyond_Binary_Attributes.pdf
_version_ 1789483258203865088