SIDNet: Learning Shading-aware Illumination Descriptor for Image Harmonization

Image harmonization aims at adjusting the appearance of the foreground to make it more compatible with the background. Without exploring background illumination and its effects on the foreground elements, existing works are incapable of generating a realistic foreground shading. In this paper, we de...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hu, Zhongyun, Nsampi, Ntumba Elie, Wang, Xue, Wang, Qing
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Hu, Zhongyun
Nsampi, Ntumba Elie
Wang, Xue
Wang, Qing
description Image harmonization aims at adjusting the appearance of the foreground to make it more compatible with the background. Without exploring background illumination and its effects on the foreground elements, existing works are incapable of generating a realistic foreground shading. In this paper, we decompose the image harmonization task into two sub-problems: 1) illumination estimation of the background image and 2) re-rendering of foreground objects under background illumination. Before solving these two sub-problems, we first learn a shading-aware illumination descriptor via a well-designed neural rendering framework, of which the key is a shading bases module that generates multiple shading bases from the foreground image. Then we design a background illumination estimation module to extract the illumination descriptor from the background. Finally, the Shading-aware Illumination Descriptor is used in conjunction with the neural rendering framework (SIDNet) to produce the harmonized foreground image containing a novel harmonized shading. Moreover, we construct a photo-realistic synthetic image harmonization dataset that contains numerous shading variations with image-based lighting. Extensive experiments on both synthetic and real data demonstrate the superiority of the proposed method, especially in dealing with foreground shadings.
doi_str_mv 10.48550/arxiv.2112.01314
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2112_01314</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2112_01314</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-b45e0185c9f7b6a9025fe710f5154eaa4e9f1a7c30c5cd1190992175da9c471c3</originalsourceid><addsrcrecordid>eNotj71OwzAURr0woMIDMOEXSPBNfOuaDbVAI0VlaPfo1rkulhKncsPv01MCw6ezfDrSEeIGVK4XiOqO0md4zwuAIldQgr4Um2212vB4L2umFEM8yO0rtWdm9EGJZdV1b32INIYhyhWfXArHcUjSn1f1dGC5ptQPMXxPlytx4ak78fU_Z2L39LhbrrP65blaPtQZzY3O9hpZwQKd9WY_J6sK9GxAeQTUTKTZeiDjSuXQtQBWWVuAwZas0wZcORO3f9opqDmm0FP6an7Dmims_AFWuUhv</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>SIDNet: Learning Shading-aware Illumination Descriptor for Image Harmonization</title><source>arXiv.org</source><creator>Hu, Zhongyun ; Nsampi, Ntumba Elie ; Wang, Xue ; Wang, Qing</creator><creatorcontrib>Hu, Zhongyun ; Nsampi, Ntumba Elie ; Wang, Xue ; Wang, Qing</creatorcontrib><description>Image harmonization aims at adjusting the appearance of the foreground to make it more compatible with the background. Without exploring background illumination and its effects on the foreground elements, existing works are incapable of generating a realistic foreground shading. In this paper, we decompose the image harmonization task into two sub-problems: 1) illumination estimation of the background image and 2) re-rendering of foreground objects under background illumination. Before solving these two sub-problems, we first learn a shading-aware illumination descriptor via a well-designed neural rendering framework, of which the key is a shading bases module that generates multiple shading bases from the foreground image. Then we design a background illumination estimation module to extract the illumination descriptor from the background. Finally, the Shading-aware Illumination Descriptor is used in conjunction with the neural rendering framework (SIDNet) to produce the harmonized foreground image containing a novel harmonized shading. Moreover, we construct a photo-realistic synthetic image harmonization dataset that contains numerous shading variations with image-based lighting. Extensive experiments on both synthetic and real data demonstrate the superiority of the proposed method, especially in dealing with foreground shadings.</description><identifier>DOI: 10.48550/arxiv.2112.01314</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2021-12</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2112.01314$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2112.01314$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Hu, Zhongyun</creatorcontrib><creatorcontrib>Nsampi, Ntumba Elie</creatorcontrib><creatorcontrib>Wang, Xue</creatorcontrib><creatorcontrib>Wang, Qing</creatorcontrib><title>SIDNet: Learning Shading-aware Illumination Descriptor for Image Harmonization</title><description>Image harmonization aims at adjusting the appearance of the foreground to make it more compatible with the background. Without exploring background illumination and its effects on the foreground elements, existing works are incapable of generating a realistic foreground shading. In this paper, we decompose the image harmonization task into two sub-problems: 1) illumination estimation of the background image and 2) re-rendering of foreground objects under background illumination. Before solving these two sub-problems, we first learn a shading-aware illumination descriptor via a well-designed neural rendering framework, of which the key is a shading bases module that generates multiple shading bases from the foreground image. Then we design a background illumination estimation module to extract the illumination descriptor from the background. Finally, the Shading-aware Illumination Descriptor is used in conjunction with the neural rendering framework (SIDNet) to produce the harmonized foreground image containing a novel harmonized shading. Moreover, we construct a photo-realistic synthetic image harmonization dataset that contains numerous shading variations with image-based lighting. Extensive experiments on both synthetic and real data demonstrate the superiority of the proposed method, especially in dealing with foreground shadings.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj71OwzAURr0woMIDMOEXSPBNfOuaDbVAI0VlaPfo1rkulhKncsPv01MCw6ezfDrSEeIGVK4XiOqO0md4zwuAIldQgr4Um2212vB4L2umFEM8yO0rtWdm9EGJZdV1b32INIYhyhWfXArHcUjSn1f1dGC5ptQPMXxPlytx4ak78fU_Z2L39LhbrrP65blaPtQZzY3O9hpZwQKd9WY_J6sK9GxAeQTUTKTZeiDjSuXQtQBWWVuAwZas0wZcORO3f9opqDmm0FP6an7Dmims_AFWuUhv</recordid><startdate>20211202</startdate><enddate>20211202</enddate><creator>Hu, Zhongyun</creator><creator>Nsampi, Ntumba Elie</creator><creator>Wang, Xue</creator><creator>Wang, Qing</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20211202</creationdate><title>SIDNet: Learning Shading-aware Illumination Descriptor for Image Harmonization</title><author>Hu, Zhongyun ; Nsampi, Ntumba Elie ; Wang, Xue ; Wang, Qing</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-b45e0185c9f7b6a9025fe710f5154eaa4e9f1a7c30c5cd1190992175da9c471c3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Hu, Zhongyun</creatorcontrib><creatorcontrib>Nsampi, Ntumba Elie</creatorcontrib><creatorcontrib>Wang, Xue</creatorcontrib><creatorcontrib>Wang, Qing</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Hu, Zhongyun</au><au>Nsampi, Ntumba Elie</au><au>Wang, Xue</au><au>Wang, Qing</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>SIDNet: Learning Shading-aware Illumination Descriptor for Image Harmonization</atitle><date>2021-12-02</date><risdate>2021</risdate><abstract>Image harmonization aims at adjusting the appearance of the foreground to make it more compatible with the background. Without exploring background illumination and its effects on the foreground elements, existing works are incapable of generating a realistic foreground shading. In this paper, we decompose the image harmonization task into two sub-problems: 1) illumination estimation of the background image and 2) re-rendering of foreground objects under background illumination. Before solving these two sub-problems, we first learn a shading-aware illumination descriptor via a well-designed neural rendering framework, of which the key is a shading bases module that generates multiple shading bases from the foreground image. Then we design a background illumination estimation module to extract the illumination descriptor from the background. Finally, the Shading-aware Illumination Descriptor is used in conjunction with the neural rendering framework (SIDNet) to produce the harmonized foreground image containing a novel harmonized shading. Moreover, we construct a photo-realistic synthetic image harmonization dataset that contains numerous shading variations with image-based lighting. Extensive experiments on both synthetic and real data demonstrate the superiority of the proposed method, especially in dealing with foreground shadings.</abstract><doi>10.48550/arxiv.2112.01314</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2112.01314
ispartof
issn
language eng
recordid cdi_arxiv_primary_2112_01314
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
title SIDNet: Learning Shading-aware Illumination Descriptor for Image Harmonization
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-27T01%3A08%3A43IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=SIDNet:%20Learning%20Shading-aware%20Illumination%20Descriptor%20for%20Image%20Harmonization&rft.au=Hu,%20Zhongyun&rft.date=2021-12-02&rft_id=info:doi/10.48550/arxiv.2112.01314&rft_dat=%3Carxiv_GOX%3E2112_01314%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true