Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation
Abstract Unsupervised heterogeneous face translation requires obtaining heterogeneous images with the same identities at training time, limiting the use in unconstrained real‐world scenarios. Taking a step further towards unconstrained heterogeneous face translation, the authors explore unsupervised...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-12-01
|
| Series: | Electronics Letters |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/ell2.13034 |
| _version_ | 1797383542400876544 |
|---|---|
| author | Zhihui Liu Jian Chen Tingshuai Liu Yinghui Zhang |
| author_facet | Zhihui Liu Jian Chen Tingshuai Liu Yinghui Zhang |
| author_sort | Zhihui Liu |
| collection | DOAJ |
| description | Abstract Unsupervised heterogeneous face translation requires obtaining heterogeneous images with the same identities at training time, limiting the use in unconstrained real‐world scenarios. Taking a step further towards unconstrained heterogeneous face translation, the authors explore unsupervised zero‐shot heterogeneous face translation for the first time, which is expected to synthesize images that resemble the style of target images and whose identities in the source domain have been preserved but never seen in the target domain during training. Essentially, asymmetry between heterogeneous faces under the zero‐shot setting further exacerbates the distortion and blurring of the translated images. The authors therefore propose a novel frequency‐structure‐guided regularization, which can jointly encourage to capture detailed textures and maintain identity consistency. Through extensive experimental validation and comparisons to several baseline methods on benchmark datasets, the authors verify the effectiveness of the proposed framework. |
| first_indexed | 2024-03-08T21:22:29Z |
| format | Article |
| id | doaj.art-ba740dc462f04d3789b10359f91076e7 |
| institution | Directory Open Access Journal |
| issn | 0013-5194 1350-911X |
| language | English |
| last_indexed | 2024-03-08T21:22:29Z |
| publishDate | 2023-12-01 |
| publisher | Wiley |
| record_format | Article |
| series | Electronics Letters |
| spelling | doaj.art-ba740dc462f04d3789b10359f91076e72023-12-21T08:55:55ZengWileyElectronics Letters0013-51941350-911X2023-12-015923n/an/a10.1049/ell2.13034Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translationZhihui Liu0Jian Chen1Tingshuai Liu2Yinghui Zhang3Space Star Technology Co., LTD Beijing ChinaSpace Star Technology Co., LTD Beijing ChinaCollege of Education for the Future Beijing Normal University Zhuhai Guangdong Province ChinaCollege of Education for the Future Beijing Normal University Zhuhai Guangdong Province ChinaAbstract Unsupervised heterogeneous face translation requires obtaining heterogeneous images with the same identities at training time, limiting the use in unconstrained real‐world scenarios. Taking a step further towards unconstrained heterogeneous face translation, the authors explore unsupervised zero‐shot heterogeneous face translation for the first time, which is expected to synthesize images that resemble the style of target images and whose identities in the source domain have been preserved but never seen in the target domain during training. Essentially, asymmetry between heterogeneous faces under the zero‐shot setting further exacerbates the distortion and blurring of the translated images. The authors therefore propose a novel frequency‐structure‐guided regularization, which can jointly encourage to capture detailed textures and maintain identity consistency. Through extensive experimental validation and comparisons to several baseline methods on benchmark datasets, the authors verify the effectiveness of the proposed framework.https://doi.org/10.1049/ell2.13034computer visionface recognitionimage enhancementimage matching |
| spellingShingle | Zhihui Liu Jian Chen Tingshuai Liu Yinghui Zhang Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation Electronics Letters computer vision face recognition image enhancement image matching |
| title | Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation |
| title_full | Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation |
| title_fullStr | Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation |
| title_full_unstemmed | Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation |
| title_short | Frequency‐structure‐aware modelling for unsupervised zero‐shot cross‐domain heterogeneous face translation |
| title_sort | frequency structure aware modelling for unsupervised zero shot cross domain heterogeneous face translation |
| topic | computer vision face recognition image enhancement image matching |
| url | https://doi.org/10.1049/ell2.13034 |
| work_keys_str_mv | AT zhihuiliu frequencystructureawaremodellingforunsupervisedzeroshotcrossdomainheterogeneousfacetranslation AT jianchen frequencystructureawaremodellingforunsupervisedzeroshotcrossdomainheterogeneousfacetranslation AT tingshuailiu frequencystructureawaremodellingforunsupervisedzeroshotcrossdomainheterogeneousfacetranslation AT yinghuizhang frequencystructureawaremodellingforunsupervisedzeroshotcrossdomainheterogeneousfacetranslation |