GH-Feat : Learning Versatile Generative Hierarchical Features From GANs

Recent years witness the tremendous success of generative adversarial networks (GANs) in synthesizing photo-realistic images. GAN generator learns to compose realistic images and reproduce the real data distribution. Through that, a hierarchical visual feature with multi-level semantics spontaneousl...

Ausführliche Beschreibung

Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on pattern analysis and machine intelligence. - 1979. - 45(2023), 6 vom: 01. Juni, Seite 7395-7411
1. Verfasser: Xu, Yinghao (VerfasserIn)
Weitere Verfasser: Shen, Yujun, Zhu, Jiapeng, Yang, Ceyuan, Zhou, Bolei
Format: Online-Aufsatz
Sprache:English
Veröffentlicht: 2023
Zugriff auf das übergeordnete Werk:IEEE transactions on pattern analysis and machine intelligence
Schlagworte:Journal Article
LEADER 01000caa a22002652c 4500
001 NLM349677158
003 DE-627
005 20250304044143.0
007 cr uuu---uuuuu
008 231226s2023 xx |||||o 00| ||eng c
024 7 |a 10.1109/TPAMI.2022.3225788  |2 doi 
028 5 2 |a pubmed25n1165.xml 
035 |a (DE-627)NLM349677158 
035 |a (NLM)36455092 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Xu, Yinghao  |e verfasserin  |4 aut 
245 1 0 |a GH-Feat  |b Learning Versatile Generative Hierarchical Features From GANs 
264 1 |c 2023 
336 |a Text  |b txt  |2 rdacontent 
337 |a ƒaComputermedien  |b c  |2 rdamedia 
338 |a ƒa Online-Ressource  |b cr  |2 rdacarrier 
500 |a Date Completed 07.05.2023 
500 |a Date Revised 07.05.2023 
500 |a published: Print-Electronic 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a Recent years witness the tremendous success of generative adversarial networks (GANs) in synthesizing photo-realistic images. GAN generator learns to compose realistic images and reproduce the real data distribution. Through that, a hierarchical visual feature with multi-level semantics spontaneously emerges. In this work we investigate that such a generative feature learned from image synthesis exhibits great potentials in solving a wide range of computer vision tasks, including both generative ones and more importantly discriminative ones. We first train an encoder by considering the pre-trained StyleGAN generator as a learned loss function. The visual features produced by our encoder, termed as Generative Hierarchical Features (GH-Feat), highly align with the layer-wise GAN representations, and hence describe the input image adequately from the reconstruction perspective. Extensive experiments support the versatile transferability of GH-Feat across a range of applications, such as image editing, image processing, image harmonization, face verification, landmark detection, layout prediction, image retrieval, etc. We further show that, through a proper spatial expansion, our developed GH-Feat can also facilitate fine-grained semantic segmentation using only a few annotations. Both qualitative and quantitative results demonstrate the appealing performance of GH-Feat. Code and models are available at https://genforce.github.io/ghfeat/ 
650 4 |a Journal Article 
700 1 |a Shen, Yujun  |e verfasserin  |4 aut 
700 1 |a Zhu, Jiapeng  |e verfasserin  |4 aut 
700 1 |a Yang, Ceyuan  |e verfasserin  |4 aut 
700 1 |a Zhou, Bolei  |e verfasserin  |4 aut 
773 0 8 |i Enthalten in  |t IEEE transactions on pattern analysis and machine intelligence  |d 1979  |g 45(2023), 6 vom: 01. Juni, Seite 7395-7411  |w (DE-627)NLM098212257  |x 1939-3539  |7 nnas 
773 1 8 |g volume:45  |g year:2023  |g number:6  |g day:01  |g month:06  |g pages:7395-7411 
856 4 0 |u http://dx.doi.org/10.1109/TPAMI.2022.3225788  |3 Volltext 
912 |a GBV_USEFLAG_A 
912 |a SYSFLAG_A 
912 |a GBV_NLM 
912 |a GBV_ILN_350 
951 |a AR 
952 |d 45  |j 2023  |e 6  |b 01  |c 06  |h 7395-7411