Generative Adversarial Shaders for Real-Time Realism Enhancement

Application of realism enhancement methods, particularly in real-time and resource-constrained settings, has been frustrated by the expense of existing methods. These achieve high quality results only at the cost of long runtimes and high bandwidth, memory, and power requirements. We present an effi...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Salmi, Arturo, Cséfalvay, Szabolcs, Imber, James
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 Salmi, Arturo
Cséfalvay, Szabolcs
Imber, James
description Application of realism enhancement methods, particularly in real-time and resource-constrained settings, has been frustrated by the expense of existing methods. These achieve high quality results only at the cost of long runtimes and high bandwidth, memory, and power requirements. We present an efficient alternative: a high-performance, generative shader-based approach that adapts machine learning techniques to real-time applications, even in resource-constrained settings such as embedded and mobile GPUs. The proposed learnable shader pipeline comprises differentiable functions that can be trained in an end-to-end manner using an adversarial objective, allowing for faithful reproduction of the appearance of a target image set without manual tuning. The shader pipeline is optimized for highly efficient execution on the target device, providing temporally stable, faster-than-real time results with quality competitive with many neural network-based methods.
doi_str_mv 10.48550/arxiv.2306.04629
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2306_04629</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2306_04629</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-aeaf957ec6f400bd018c2ae28aa6087b78e2120172fdf8f6353f3a8e0e2570e03</originalsourceid><addsrcrecordid>eNotz81qg0AUBeDZZFGSPkBXnRfQXGecH3cVsbYQCLTu5Ubv4IDaMoqkb9_EdnXO2Rz4GHtKIE6tUnDEcPVrLCToGFItsgf2UtFEARe_Es-7lcKMwePAP3vsboO7r8A_CIeo9iNtzc8jL6cep5ZGmpYD2zkcZnr8zz2rX8u6eItO5-q9yE8RapNFSOgyZajVLgW4dJDYViAJi6jBmouxJBIBiRGuc9ZpqaSTaAlIKAMEcs-e_243QvMd_Ijhp7lTmo0ifwEOMUOu</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Generative Adversarial Shaders for Real-Time Realism Enhancement</title><source>arXiv.org</source><creator>Salmi, Arturo ; Cséfalvay, Szabolcs ; Imber, James</creator><creatorcontrib>Salmi, Arturo ; Cséfalvay, Szabolcs ; Imber, James</creatorcontrib><description>Application of realism enhancement methods, particularly in real-time and resource-constrained settings, has been frustrated by the expense of existing methods. These achieve high quality results only at the cost of long runtimes and high bandwidth, memory, and power requirements. We present an efficient alternative: a high-performance, generative shader-based approach that adapts machine learning techniques to real-time applications, even in resource-constrained settings such as embedded and mobile GPUs. The proposed learnable shader pipeline comprises differentiable functions that can be trained in an end-to-end manner using an adversarial objective, allowing for faithful reproduction of the appearance of a target image set without manual tuning. The shader pipeline is optimized for highly efficient execution on the target device, providing temporally stable, faster-than-real time results with quality competitive with many neural network-based methods.</description><identifier>DOI: 10.48550/arxiv.2306.04629</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition ; Computer Science - Graphics ; Computer Science - Learning</subject><creationdate>2023-05</creationdate><rights>http://creativecommons.org/licenses/by-nc-nd/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2306.04629$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2306.04629$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Salmi, Arturo</creatorcontrib><creatorcontrib>Cséfalvay, Szabolcs</creatorcontrib><creatorcontrib>Imber, James</creatorcontrib><title>Generative Adversarial Shaders for Real-Time Realism Enhancement</title><description>Application of realism enhancement methods, particularly in real-time and resource-constrained settings, has been frustrated by the expense of existing methods. These achieve high quality results only at the cost of long runtimes and high bandwidth, memory, and power requirements. We present an efficient alternative: a high-performance, generative shader-based approach that adapts machine learning techniques to real-time applications, even in resource-constrained settings such as embedded and mobile GPUs. The proposed learnable shader pipeline comprises differentiable functions that can be trained in an end-to-end manner using an adversarial objective, allowing for faithful reproduction of the appearance of a target image set without manual tuning. The shader pipeline is optimized for highly efficient execution on the target device, providing temporally stable, faster-than-real time results with quality competitive with many neural network-based methods.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Computer Science - Graphics</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz81qg0AUBeDZZFGSPkBXnRfQXGecH3cVsbYQCLTu5Ubv4IDaMoqkb9_EdnXO2Rz4GHtKIE6tUnDEcPVrLCToGFItsgf2UtFEARe_Es-7lcKMwePAP3vsboO7r8A_CIeo9iNtzc8jL6cep5ZGmpYD2zkcZnr8zz2rX8u6eItO5-q9yE8RapNFSOgyZajVLgW4dJDYViAJi6jBmouxJBIBiRGuc9ZpqaSTaAlIKAMEcs-e_243QvMd_Ijhp7lTmo0ifwEOMUOu</recordid><startdate>20230524</startdate><enddate>20230524</enddate><creator>Salmi, Arturo</creator><creator>Cséfalvay, Szabolcs</creator><creator>Imber, James</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20230524</creationdate><title>Generative Adversarial Shaders for Real-Time Realism Enhancement</title><author>Salmi, Arturo ; Cséfalvay, Szabolcs ; Imber, James</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-aeaf957ec6f400bd018c2ae28aa6087b78e2120172fdf8f6353f3a8e0e2570e03</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Computer Science - Graphics</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Salmi, Arturo</creatorcontrib><creatorcontrib>Cséfalvay, Szabolcs</creatorcontrib><creatorcontrib>Imber, James</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Salmi, Arturo</au><au>Cséfalvay, Szabolcs</au><au>Imber, James</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Generative Adversarial Shaders for Real-Time Realism Enhancement</atitle><date>2023-05-24</date><risdate>2023</risdate><abstract>Application of realism enhancement methods, particularly in real-time and resource-constrained settings, has been frustrated by the expense of existing methods. These achieve high quality results only at the cost of long runtimes and high bandwidth, memory, and power requirements. We present an efficient alternative: a high-performance, generative shader-based approach that adapts machine learning techniques to real-time applications, even in resource-constrained settings such as embedded and mobile GPUs. The proposed learnable shader pipeline comprises differentiable functions that can be trained in an end-to-end manner using an adversarial objective, allowing for faithful reproduction of the appearance of a target image set without manual tuning. The shader pipeline is optimized for highly efficient execution on the target device, providing temporally stable, faster-than-real time results with quality competitive with many neural network-based methods.</abstract><doi>10.48550/arxiv.2306.04629</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2306.04629
ispartof
issn
language eng
recordid cdi_arxiv_primary_2306_04629
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
Computer Science - Graphics
Computer Science - Learning
title Generative Adversarial Shaders for Real-Time Realism Enhancement
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-01T03%3A18%3A36IST&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=Generative%20Adversarial%20Shaders%20for%20Real-Time%20Realism%20Enhancement&rft.au=Salmi,%20Arturo&rft.date=2023-05-24&rft_id=info:doi/10.48550/arxiv.2306.04629&rft_dat=%3Carxiv_GOX%3E2306_04629%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