DETR with Additional Global Aggregation for Cross-domain Weakly Supervised Object Detection

Tang, Z; Sun, Y; Liu, S; Yang, Y

DETR with Additional Global Aggregation for Cross-domain Weakly Supervised Object Detection

Tang, Z Sun, Y Liu, S Yang, Y

Permalink

Publisher:: IEEE
Publication Type:: Conference Proceeding
Citation:: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023, 2023-June, pp. 11422-11432
Issue Date:: 2023-01-01

Closed Access

	Filename	Description	Size
	2304.07082v1.pdf	Published version	1.4 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Tang, Z
dc.contributor.author	Sun, Y
dc.contributor.author	Liu, S
dc.contributor.author	Yang, Y https://orcid.org/0000-0002-0512-880X
dc.date	2023-06-17
dc.date.accessioned	2024-06-12T20:17:11Z
dc.date.available	2024-06-12T20:17:11Z
dc.date.issued	2023-01-01
dc.identifier.citation	2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023, 2023-June, pp. 11422-11432
dc.identifier.issn	1063-6919
dc.identifier.uri	http://hdl.handle.net/10453/179499
dc.description.abstract	This paper presents a DETR based method for cross domain weakly supervised object detection CDWSOD aiming at adapting the detector from source to target domain through weak supervision We think DETR has strong potential for CDWSOD due to an insight the encoder and the decoder in DETR are both based on the attention mechanism and are thus capable of aggregating semantics across the entire image The aggregation results i e image level predictions can naturally exploit the weak supervision for domain alignment Such motivated we propose DETR with additional Global Aggregation DETR GA a CDWSOD detector that simultaneously makes instance level image level predictions and utilizes strong weak supervisions The key point of DETR GA is very simple for the encoder decoder we respectively add multiple class queries a foreground query to aggregate the semantics into image level predictions Our query based aggregation has two advantages First in the encoder the weakly supervised class queries are capable of roughly locating the corresponding positions and excluding the distraction from non relevant regions Second through our design the object queries and the foreground query in the decoder share consensus on the class semantics therefore making the strong and weak supervision mutually benefit each other for domain alignment Extensive experiments on four popular cross domain benchmarks show that DETR GA significantly improves cross domain detection accuracy e g 29 0 79 4 mAP on PASCAL VOC Clipartall dataset and advances the states of the art
dc.language	en
dc.publisher	IEEE
dc.relation.ispartof	2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
dc.relation.ispartof	2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition
dc.relation.isbasedon	10.1109/cvpr52729.2023.01099
dc.rights	info:eu-repo/semantics/closedAccess
dc.title	DETR with Additional Global Aggregation for Cross-domain Weakly Supervised Object Detection
dc.type	Conference Proceeding
utslib.citation.volume	2023-June
utslib.location.activity	Vancouver, BC, Canada
pubs.organisational-group	University of Technology Sydney
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology
utslib.copyright.status	closed_access	*
dc.date.updated	2024-06-12T20:17:09Z
pubs.finish-date	2023-06-24
pubs.publication-status	Published
pubs.start-date	2023-06-17
pubs.volume	2023-June

Abstract:

This paper presents a DETR based method for cross domain weakly supervised object detection CDWSOD aiming at adapting the detector from source to target domain through weak supervision We think DETR has strong potential for CDWSOD due to an insight the encoder and the decoder in DETR are both based on the attention mechanism and are thus capable of aggregating semantics across the entire image The aggregation results i e image level predictions can naturally exploit the weak supervision for domain alignment Such motivated we propose DETR with additional Global Aggregation DETR GA a CDWSOD detector that simultaneously makes instance level image level predictions and utilizes strong weak supervisions The key point of DETR GA is very simple for the encoder decoder we respectively add multiple class queries a foreground query to aggregate the semantics into image level predictions Our query based aggregation has two advantages First in the encoder the weakly supervised class queries are capable of roughly locating the corresponding positions and excluding the distraction from non relevant regions Second through our design the object queries and the foreground query in the decoder share consensus on the class semantics therefore making the strong and weak supervision mutually benefit each other for domain alignment Extensive experiments on four popular cross domain benchmarks show that DETR GA significantly improves cross domain detection accuracy e g 29 0 79 4 mAP on PASCAL VOC Clipartall dataset and advances the states of the art

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/179499