RAID: a relation-augmented image descriptor

Handle URI:
http://hdl.handle.net/10754/620947
Title:
RAID: a relation-augmented image descriptor
Authors:
Guerrero, Paul; Mitra, Niloy J.; Wonka, Peter ( 0000-0003-0627-9746 )
Abstract:
As humans, we regularly interpret scenes based on how objects are related, rather than based on the objects themselves. For example, we see a person riding an object X or a plank bridging two objects. Current methods provide limited support to search for content based on such relations. We present RAID, a relation-augmented image descriptor that supports queries based on inter-region relations. The key idea of our descriptor is to encode region-to-region relations as the spatial distribution of point-to-region relationships between two image regions. RAID allows sketch-based retrieval and requires minimal training data, thus making it suited even for querying uncommon relations. We evaluate the proposed descriptor by querying into large image databases and successfully extract nontrivial images demonstrating complex inter-region relations, which are easily missed or erroneously classified by existing methods. We assess the robustness of RAID on multiple datasets even when the region segmentation is computed automatically or very noisy.
KAUST Department:
CEMSE; VCC
Citation:
Guerrero P, Mitra NJ, Wonka P (2016) RAID. ACM Transactions on Graphics 35: 1–12. Available: http://dx.doi.org/10.1145/2897824.2925939.
Publisher:
Association for Computing Machinery (ACM)
KAUST Grant Number:
OCRF-2014-CGR3-62140401
Conference/Event name:
ACM SIGGRAPH 2016
Issue Date:
11-Jul-2016
DOI:
10.1145/2897824.2925939
Type:
Conference Paper
ISSN:
0730-0301
Sponsors:
The research described here was supported by the Office of Sponsored Research (OSR) under Award No. OCRF-2014-CGR3-62140401, the Visual Computing Center at KAUST, ERC Starting Grant SmartGeometry (StG-2013 335373), Marie Curie CIG 303541 and the Open3D Project (EPSRC Grant EP/M013685/1).
Additional Links:
http://dl.acm.org/citation.cfm?doid=2897824.2925939
Appears in Collections:
Conference Papers

Full metadata record

DC FieldValue Language
dc.contributor.authorGuerrero, Paulen
dc.contributor.authorMitra, Niloy J.en
dc.contributor.authorWonka, Peteren
dc.date.accessioned2016-10-12T09:15:28Z-
dc.date.available2016-10-12T09:15:28Z-
dc.date.issued2016-07-11en
dc.identifier.citationGuerrero P, Mitra NJ, Wonka P (2016) RAID. ACM Transactions on Graphics 35: 1–12. Available: http://dx.doi.org/10.1145/2897824.2925939.en
dc.identifier.issn0730-0301en
dc.identifier.doi10.1145/2897824.2925939en
dc.identifier.urihttp://hdl.handle.net/10754/620947-
dc.description.abstractAs humans, we regularly interpret scenes based on how objects are related, rather than based on the objects themselves. For example, we see a person riding an object X or a plank bridging two objects. Current methods provide limited support to search for content based on such relations. We present RAID, a relation-augmented image descriptor that supports queries based on inter-region relations. The key idea of our descriptor is to encode region-to-region relations as the spatial distribution of point-to-region relationships between two image regions. RAID allows sketch-based retrieval and requires minimal training data, thus making it suited even for querying uncommon relations. We evaluate the proposed descriptor by querying into large image databases and successfully extract nontrivial images demonstrating complex inter-region relations, which are easily missed or erroneously classified by existing methods. We assess the robustness of RAID on multiple datasets even when the region segmentation is computed automatically or very noisy.en
dc.description.sponsorshipThe research described here was supported by the Office of Sponsored Research (OSR) under Award No. OCRF-2014-CGR3-62140401, the Visual Computing Center at KAUST, ERC Starting Grant SmartGeometry (StG-2013 335373), Marie Curie CIG 303541 and the Open3D Project (EPSRC Grant EP/M013685/1).en
dc.publisherAssociation for Computing Machinery (ACM)en
dc.relation.urlhttp://dl.acm.org/citation.cfm?doid=2897824.2925939en
dc.rights(c) 2016 Copyright held by the owner/author(s). This work in licensed under Creative Commons Attribution International 4.0 license.en
dc.subjectImage descriptorsen
dc.titleRAID: a relation-augmented image descriptoren
dc.typeConference Paperen
dc.contributor.departmentCEMSEen
dc.contributor.departmentVCCen
dc.conference.date2016-07-24 to 2016-07-28en
dc.conference.nameACM SIGGRAPH 2016en
dc.conference.locationAnaheim, CA, USAen
dc.eprint.versionPublisher's Version/PDFen
dc.contributor.institutionUniversity College Londonen
kaust.authorGuerrero, Paulen
kaust.authorWonka, Peteren
kaust.grant.numberOCRF-2014-CGR3-62140401en
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.