The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Devido ao surto global de coronavírus, as pessoas usam cada vez mais máscaras, mesmo quando fotografadas. Como resultado, as fotos carregadas em páginas web e serviços de redes sociais com a metade inferior do rosto oculta têm menos probabilidade de transmitir a atratividade das pessoas fotografadas. Neste estudo, propomos um método para completar regiões de máscara facial usando StyleGAN2, um tipo de Rede Adversarial Generativa (GAN). No método proposto, uma imagem de referência da mesma pessoa sem máscara é preparada separadamente de uma imagem alvo da pessoa usando máscara. Depois que a região da máscara na imagem alvo é temporariamente pintada, a orientação do rosto e o contorno da pessoa na imagem de referência são alterados para corresponder aos da imagem alvo usando StyleGAN2. A imagem alterada é então composta na região da máscara enquanto se corrige o tom da cor para produzir uma imagem sem máscara, preservando as características da pessoa.
Norihiko KAWAI
Osaka Institute of Technology
Hiroaki KOIKE
Osaka Institute of Technology
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Norihiko KAWAI, Hiroaki KOIKE, "Facial Mask Completion Using StyleGAN2 Preserving Features of the Person" in IEICE TRANSACTIONS on Information,
vol. E106-D, no. 10, pp. 1627-1637, October 2023, doi: 10.1587/transinf.2023PCP0002.
Abstract: Due to the global outbreak of coronaviruses, people are increasingly wearing masks even when photographed. As a result, photos uploaded to web pages and social networking services with the lower half of the face hidden are less likely to convey the attractiveness of the photographed persons. In this study, we propose a method to complete facial mask regions using StyleGAN2, a type of Generative Adversarial Networks (GAN). In the proposed method, a reference image of the same person without a mask is prepared separately from a target image of the person wearing a mask. After the mask region in the target image is temporarily inpainted, the face orientation and contour of the person in the reference image are changed to match those of the target image using StyleGAN2. The changed image is then composited into the mask region while correcting the color tone to produce a mask-free image while preserving the person's features.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2023PCP0002/_p
Copiar
@ARTICLE{e106-d_10_1627,
author={Norihiko KAWAI, Hiroaki KOIKE, },
journal={IEICE TRANSACTIONS on Information},
title={Facial Mask Completion Using StyleGAN2 Preserving Features of the Person},
year={2023},
volume={E106-D},
number={10},
pages={1627-1637},
abstract={Due to the global outbreak of coronaviruses, people are increasingly wearing masks even when photographed. As a result, photos uploaded to web pages and social networking services with the lower half of the face hidden are less likely to convey the attractiveness of the photographed persons. In this study, we propose a method to complete facial mask regions using StyleGAN2, a type of Generative Adversarial Networks (GAN). In the proposed method, a reference image of the same person without a mask is prepared separately from a target image of the person wearing a mask. After the mask region in the target image is temporarily inpainted, the face orientation and contour of the person in the reference image are changed to match those of the target image using StyleGAN2. The changed image is then composited into the mask region while correcting the color tone to produce a mask-free image while preserving the person's features.},
keywords={},
doi={10.1587/transinf.2023PCP0002},
ISSN={1745-1361},
month={October},}
Copiar
TY - JOUR
TI - Facial Mask Completion Using StyleGAN2 Preserving Features of the Person
T2 - IEICE TRANSACTIONS on Information
SP - 1627
EP - 1637
AU - Norihiko KAWAI
AU - Hiroaki KOIKE
PY - 2023
DO - 10.1587/transinf.2023PCP0002
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E106-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2023
AB - Due to the global outbreak of coronaviruses, people are increasingly wearing masks even when photographed. As a result, photos uploaded to web pages and social networking services with the lower half of the face hidden are less likely to convey the attractiveness of the photographed persons. In this study, we propose a method to complete facial mask regions using StyleGAN2, a type of Generative Adversarial Networks (GAN). In the proposed method, a reference image of the same person without a mask is prepared separately from a target image of the person wearing a mask. After the mask region in the target image is temporarily inpainted, the face orientation and contour of the person in the reference image are changed to match those of the target image using StyleGAN2. The changed image is then composited into the mask region while correcting the color tone to produce a mask-free image while preserving the person's features.
ER -