Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting

Recently, context learning networks have shown promise in filling large holes in natural images. These networks can decorate the predicted contents with high-frequency details by borrowing or copying neural information from the known region. However, this operation might introduce undesired content...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Liang Liao, Ruimin Hu, Jing Xiao, Zhongyuan Wang
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2019
Materias:	Neural network disentangled representation image inpainting content inference style decoration Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/d2d758844a1b48b6a09a7d4b1ef2b6a2
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:d2d758844a1b48b6a09a7d4b1ef2b6a2
record_format	dspace
spelling	oai:doaj.org-article:d2d758844a1b48b6a09a7d4b1ef2b6a22021-11-19T00:02:22ZArtist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting2169-353610.1109/ACCESS.2019.2905268https://doaj.org/article/d2d758844a1b48b6a09a7d4b1ef2b6a22019-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/8669751/https://doaj.org/toc/2169-3536Recently, context learning networks have shown promise in filling large holes in natural images. These networks can decorate the predicted contents with high-frequency details by borrowing or copying neural information from the known region. However, this operation might introduce undesired content change in the synthesized region, especially when similar neural patterns cannot be found in the known region. To solve this problem, we present a network named Artist-Net to decompose an image into the content code and style code explicitly. The Artist-Net completes a corrupted image following the way an artist restores a damaged picture. It can produce more detailed content by inferring the content code of the corrupted images in the latent space since the dimension of the content space is lower than the original image. The Artist-Net can also keep style consistent over the entire image by decorating the inferred content code with the style code extracted from the known region. The experiments on multiple datasets, including structural and natural images demonstrate that the proposed network out-performs the existing ones in terms of content accuracy as well as texture details.Liang LiaoRuimin HuJing XiaoZhongyuan WangIEEEarticleNeural networkdisentangled representationimage inpaintingcontent inferencestyle decorationElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 7, Pp 36921-36933 (2019)
institution	DOAJ
collection	DOAJ
language	EN
topic	Neural network disentangled representation image inpainting content inference style decoration Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Neural network disentangled representation image inpainting content inference style decoration Electrical engineering. Electronics. Nuclear engineering TK1-9971 Liang Liao Ruimin Hu Jing Xiao Zhongyuan Wang Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting
description	Recently, context learning networks have shown promise in filling large holes in natural images. These networks can decorate the predicted contents with high-frequency details by borrowing or copying neural information from the known region. However, this operation might introduce undesired content change in the synthesized region, especially when similar neural patterns cannot be found in the known region. To solve this problem, we present a network named Artist-Net to decompose an image into the content code and style code explicitly. The Artist-Net completes a corrupted image following the way an artist restores a damaged picture. It can produce more detailed content by inferring the content code of the corrupted images in the latent space since the dimension of the content space is lower than the original image. The Artist-Net can also keep style consistent over the entire image by decorating the inferred content code with the style code extracted from the known region. The experiments on multiple datasets, including structural and natural images demonstrate that the proposed network out-performs the existing ones in terms of content accuracy as well as texture details.
format	article
author	Liang Liao Ruimin Hu Jing Xiao Zhongyuan Wang
author_facet	Liang Liao Ruimin Hu Jing Xiao Zhongyuan Wang
author_sort	Liang Liao
title	Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting
title_short	Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting
title_full	Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting
title_fullStr	Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting
title_full_unstemmed	Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting
title_sort	artist-net: decorating the inferred content with unified style for image inpainting
publisher	IEEE
publishDate	2019
url	https://doaj.org/article/d2d758844a1b48b6a09a7d4b1ef2b6a2
work_keys_str_mv	AT liangliao artistnetdecoratingtheinferredcontentwithunifiedstyleforimageinpainting AT ruiminhu artistnetdecoratingtheinferredcontentwithunifiedstyleforimageinpainting AT jingxiao artistnetdecoratingtheinferredcontentwithunifiedstyleforimageinpainting AT zhongyuanwang artistnetdecoratingtheinferredcontentwithunifiedstyleforimageinpainting
_version_	1718420696795709440

Artist-Net: Decorating the Inferred Content With Unified Style for Image Inpainting

Ejemplares similares