No Cover Image

Conference Paper/Proceeding/Abstract 1179 views 136 downloads

A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images

Joss Whittle, Mark Jones Orcid Logo

Computer Graphics and Visual Computing (CGVC) 2018

Swansea University Author: Mark Jones Orcid Logo

DOI (Published version): 10.2312/cgvc.20181204

Abstract

In Full-Reference Image Quality Assessment (FR-IQA) images are compared with ground truth images that are known to be of high visual quality. These metrics are utilized in order to rank algorithms under test on their image quality performance. Throughout the progress of Monte Carlo rendering process...

Full description

Published in: Computer Graphics and Visual Computing (CGVC) 2018
Published: 2018
URI: https://cronfa.swan.ac.uk/Record/cronfa43567
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2018-08-25T19:48:25Z
last_indexed 2022-06-15T02:58:33Z
id cronfa43567
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2022-06-14T14:11:05.4794401</datestamp><bib-version>v2</bib-version><id>43567</id><entry>2018-08-25</entry><title>A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images</title><swanseaauthors><author><sid>2e1030b6e14fc9debd5d5ae7cc335562</sid><ORCID>0000-0001-8991-1190</ORCID><firstname>Mark</firstname><surname>Jones</surname><name>Mark Jones</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2018-08-25</date><deptcode>SCS</deptcode><abstract>In Full-Reference Image Quality Assessment (FR-IQA) images are compared with ground truth images that are known to be of high visual quality. These metrics are utilized in order to rank algorithms under test on their image quality performance. Throughout the progress of Monte Carlo rendering processes we often wish to determine whether images being rendered are of sufficient visual quality, without the availability of a ground truth image. In such cases FR-IQA metrics are not applicable and we instead must utilise No-Reference Image Quality Assessment (NR-IQA) measures to make predictions about the perceived quality of unconverged images. In this work we propose a deep learning approach to NR-IQA, trained specifically on noise from Monte Carlo rendering processes, which significantly outperforms existing NR-IQA methods and can produce quality predictions consistent with FR-IQA measures that have access to ground truth images.</abstract><type>Conference Paper/Proceeding/Abstract</type><journal>Computer Graphics and Visual Computing (CGVC) 2018</journal><volume/><journalNumber/><paginationStart/><paginationEnd/><publisher/><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic/><keywords/><publishedDay>13</publishedDay><publishedMonth>9</publishedMonth><publishedYear>2018</publishedYear><publishedDate>2018-09-13</publishedDate><doi>10.2312/cgvc.20181204</doi><url/><notes/><college>COLLEGE NANME</college><department>Computer Science</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>SCS</DepartmentCode><institution>Swansea University</institution><apcterm>Not Required</apcterm><lastEdited>2022-06-14T14:11:05.4794401</lastEdited><Created>2018-08-25T19:22:21.4133529</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>Joss</firstname><surname>Whittle</surname><order>1</order></author><author><firstname>Mark</firstname><surname>Jones</surname><orcid>0000-0001-8991-1190</orcid><order>2</order></author></authors><documents><document><filename>0043567-25082018192453.pdf</filename><originalFilename>2018_MC_IQA_Deep_ML.pdf</originalFilename><uploaded>2018-08-25T19:24:53.8730000</uploaded><type>Output</type><contentLength>10968717</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2019-09-13T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2022-06-14T14:11:05.4794401 v2 43567 2018-08-25 A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images 2e1030b6e14fc9debd5d5ae7cc335562 0000-0001-8991-1190 Mark Jones Mark Jones true false 2018-08-25 SCS In Full-Reference Image Quality Assessment (FR-IQA) images are compared with ground truth images that are known to be of high visual quality. These metrics are utilized in order to rank algorithms under test on their image quality performance. Throughout the progress of Monte Carlo rendering processes we often wish to determine whether images being rendered are of sufficient visual quality, without the availability of a ground truth image. In such cases FR-IQA metrics are not applicable and we instead must utilise No-Reference Image Quality Assessment (NR-IQA) measures to make predictions about the perceived quality of unconverged images. In this work we propose a deep learning approach to NR-IQA, trained specifically on noise from Monte Carlo rendering processes, which significantly outperforms existing NR-IQA methods and can produce quality predictions consistent with FR-IQA measures that have access to ground truth images. Conference Paper/Proceeding/Abstract Computer Graphics and Visual Computing (CGVC) 2018 13 9 2018 2018-09-13 10.2312/cgvc.20181204 COLLEGE NANME Computer Science COLLEGE CODE SCS Swansea University Not Required 2022-06-14T14:11:05.4794401 2018-08-25T19:22:21.4133529 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Joss Whittle 1 Mark Jones 0000-0001-8991-1190 2 0043567-25082018192453.pdf 2018_MC_IQA_Deep_ML.pdf 2018-08-25T19:24:53.8730000 Output 10968717 application/pdf Accepted Manuscript true 2019-09-13T00:00:00.0000000 true eng
title A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
spellingShingle A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
Mark Jones
title_short A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
title_full A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
title_fullStr A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
title_full_unstemmed A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
title_sort A Deep Learning Approach to No-Reference Image Quality Assessment For Monte Carlo Rendered Images
author_id_str_mv 2e1030b6e14fc9debd5d5ae7cc335562
author_id_fullname_str_mv 2e1030b6e14fc9debd5d5ae7cc335562_***_Mark Jones
author Mark Jones
author2 Joss Whittle
Mark Jones
format Conference Paper/Proceeding/Abstract
container_title Computer Graphics and Visual Computing (CGVC) 2018
publishDate 2018
institution Swansea University
doi_str_mv 10.2312/cgvc.20181204
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 1
active_str 0
description In Full-Reference Image Quality Assessment (FR-IQA) images are compared with ground truth images that are known to be of high visual quality. These metrics are utilized in order to rank algorithms under test on their image quality performance. Throughout the progress of Monte Carlo rendering processes we often wish to determine whether images being rendered are of sufficient visual quality, without the availability of a ground truth image. In such cases FR-IQA metrics are not applicable and we instead must utilise No-Reference Image Quality Assessment (NR-IQA) measures to make predictions about the perceived quality of unconverged images. In this work we propose a deep learning approach to NR-IQA, trained specifically on noise from Monte Carlo rendering processes, which significantly outperforms existing NR-IQA methods and can produce quality predictions consistent with FR-IQA measures that have access to ground truth images.
published_date 2018-09-13T03:54:48Z
_version_ 1763752740151885824
score 11.037056

Similar Items