Transformer-Based GAN for New Hairstyle Generative Networks
Traditional GAN-based image generation networks cannot accurately and naturally fuse surrounding features in local image generation tasks, especially in hairstyle generation tasks. To this end, we propose a novel transformer-based GAN for new hairstyle generation networks. The network framework comp...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/11/13/2106 |
| _version_ | 1797480368005185536 |
|---|---|
| author | Qiaoyue Man Young-Im Cho Seong-Geun Jang Hae-Jeung Lee |
| author_facet | Qiaoyue Man Young-Im Cho Seong-Geun Jang Hae-Jeung Lee |
| author_sort | Qiaoyue Man |
| collection | DOAJ |
| description | Traditional GAN-based image generation networks cannot accurately and naturally fuse surrounding features in local image generation tasks, especially in hairstyle generation tasks. To this end, we propose a novel transformer-based GAN for new hairstyle generation networks. The network framework comprises two modules: Face segmentation (F) and Transformer Generative Hairstyle (TGH) modules. The F module is used for the detection of facial and hairstyle features and the extraction of global feature masks and facial feature maps. In the TGH module, we design a transformer-based GAN to generate hairstyles and fix the details of the fusion part of faces and hairstyles in the new hairstyle generation process. To verify the effectiveness of our model, CelebA-HQ (Large-scale CelebFaces Attribute) and FFHQ (Flickr-Faces-HQ) are adopted to train and test our proposed model. In the image evaluation test used, FID, PSNR, and SSIM image evaluation methods are used to test our model and compare it with other excellent image generation networks. Our proposed model is more robust in terms of test scores and real image generation. |
| first_indexed | 2024-03-09T21:58:49Z |
| format | Article |
| id | doaj.art-9693eacaa89e4d3ea5588825d5495c60 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-09T21:58:49Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-9693eacaa89e4d3ea5588825d5495c602023-11-23T19:52:55ZengMDPI AGElectronics2079-92922022-07-011113210610.3390/electronics11132106Transformer-Based GAN for New Hairstyle Generative NetworksQiaoyue Man0Young-Im Cho1Seong-Geun Jang2Hae-Jeung Lee3Department of Computer Engineering, Gachon University, 1342 Seongnamdaero, Sujeong-gu, Seongnam-si 461-701, KoreaDepartment of Computer Engineering, Gachon University, 1342 Seongnamdaero, Sujeong-gu, Seongnam-si 461-701, KoreaDepartment of Computer Engineering, Gachon University, 1342 Seongnamdaero, Sujeong-gu, Seongnam-si 461-701, KoreaDepartment of Food & Nutrition, College of Bionano Technology, Gachon University, 1342 Seong-namdaero, Sujeong-gu, Seongnam-si 13120, KoreaTraditional GAN-based image generation networks cannot accurately and naturally fuse surrounding features in local image generation tasks, especially in hairstyle generation tasks. To this end, we propose a novel transformer-based GAN for new hairstyle generation networks. The network framework comprises two modules: Face segmentation (F) and Transformer Generative Hairstyle (TGH) modules. The F module is used for the detection of facial and hairstyle features and the extraction of global feature masks and facial feature maps. In the TGH module, we design a transformer-based GAN to generate hairstyles and fix the details of the fusion part of faces and hairstyles in the new hairstyle generation process. To verify the effectiveness of our model, CelebA-HQ (Large-scale CelebFaces Attribute) and FFHQ (Flickr-Faces-HQ) are adopted to train and test our proposed model. In the image evaluation test used, FID, PSNR, and SSIM image evaluation methods are used to test our model and compare it with other excellent image generation networks. Our proposed model is more robust in terms of test scores and real image generation.https://www.mdpi.com/2079-9292/11/13/2106face detectionconvolutional neural networkgenerative adversarial networkstransformerimage fusion |
| spellingShingle | Qiaoyue Man Young-Im Cho Seong-Geun Jang Hae-Jeung Lee Transformer-Based GAN for New Hairstyle Generative Networks Electronics face detection convolutional neural network generative adversarial networks transformer image fusion |
| title | Transformer-Based GAN for New Hairstyle Generative Networks |
| title_full | Transformer-Based GAN for New Hairstyle Generative Networks |
| title_fullStr | Transformer-Based GAN for New Hairstyle Generative Networks |
| title_full_unstemmed | Transformer-Based GAN for New Hairstyle Generative Networks |
| title_short | Transformer-Based GAN for New Hairstyle Generative Networks |
| title_sort | transformer based gan for new hairstyle generative networks |
| topic | face detection convolutional neural network generative adversarial networks transformer image fusion |
| url | https://www.mdpi.com/2079-9292/11/13/2106 |
| work_keys_str_mv | AT qiaoyueman transformerbasedganfornewhairstylegenerativenetworks AT youngimcho transformerbasedganfornewhairstylegenerativenetworks AT seonggeunjang transformerbasedganfornewhairstylegenerativenetworks AT haejeunglee transformerbasedganfornewhairstylegenerativenetworks |