Active Object Localization with Deep Reinforcement Learning

We present an active detection model for localizing objects in scenes. The model is class-specific and allows an agent to focus attention on candidate regions for identifying the correct location of a target object. This agent learns to deform a bounding box using simple transformation actions, with...

Full description

Saved in:
Bibliographic Details
Main Authors: Caicedo, Juan C., Lazebnik, Svetlana
Format: Conference Proceeding
Language:English
Subjects:
Computational modeling
Computer vision
Conferences
History
Image detection
Learning
Learning (artificial intelligence)
Localization
Position (location)
Prediction algorithms
Proposals
Reinforcement
Search problems
Transformations
Transforms
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page 2496
container_issue
container_start_page 2488
container_title
container_volume
creator Caicedo, Juan C.
Lazebnik, Svetlana
description We present an active detection model for localizing objects in scenes. The model is class-specific and allows an agent to focus attention on candidate regions for identifying the correct location of a target object. This agent learns to deform a bounding box using simple transformation actions, with the goal of determining the most specific location of target objects following top-down reasoning. The proposed localization agent is trained using deep reinforcement learning, and evaluated on the Pascal VOC 2007 dataset. We show that agents guided by the proposed model are able to localize a single instance of an object after analyzing only between 11 and 25 regions in an image, and obtain the best detection results among systems that do not use object proposals for object localization.
doi_str_mv 10.1109/ICCV.2015.286
format conference_proceeding
fullrecord <record><control><sourceid>proquest_CHZPO</sourceid><recordid>TN_cdi_proquest_miscellaneous_1816069133</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>7410643</ieee_id><sourcerecordid>1816069133</sourcerecordid><originalsourceid>FETCH-LOGICAL-i247t-35d1501455fe9b0cf706cde008b1a49eaa6dadb53c8e3bd02800585735e620593</originalsourceid><addsrcrecordid>eNotzLFLw0AUgPFTEGyro5NLRpfU9_LuLnc4lVi1ECiIuobL5UWvpElNUkX_egt1-pYfnxBXCHNEsLerLHubJ4Bqnhh9IqYodUqGLMKpmCRkIE4VyHMxHYYNANmDmoi7hR_DF0frcsN-jPLOuyb8ujF0bfQdxo_onnkXPXNo6673vOX2gNj1bWjfL8RZ7ZqBL_87E68Py5fsKc7Xj6tskcchkekYk6pQAUqlarYl-DoF7SsGMCU6adk5XbmqVOQNU1lBYgCUUSkp1gkoSzNxc_zu-u5zz8NYbMPguWlcy91-KNCgBm2R6ECvjzQwc7Hrw9b1P0UqEbQk-gOjtFN5</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype><pqid>1816069133</pqid></control><display><type>conference_proceeding</type><title>Active Object Localization with Deep Reinforcement Learning</title><source>IEEE Xplore All Conference Series</source><creator>Caicedo, Juan C. ; Lazebnik, Svetlana</creator><creatorcontrib>Caicedo, Juan C. ; Lazebnik, Svetlana</creatorcontrib><description>We present an active detection model for localizing objects in scenes. The model is class-specific and allows an agent to focus attention on candidate regions for identifying the correct location of a target object. This agent learns to deform a bounding box using simple transformation actions, with the goal of determining the most specific location of target objects following top-down reasoning. The proposed localization agent is trained using deep reinforcement learning, and evaluated on the Pascal VOC 2007 dataset. We show that agents guided by the proposed model are able to localize a single instance of an object after analyzing only between 11 and 25 regions in an image, and obtain the best detection results among systems that do not use object proposals for object localization.</description><identifier>EISSN: 2380-7504</identifier><identifier>EISBN: 1467383910</identifier><identifier>EISBN: 9781467383912</identifier><identifier>DOI: 10.1109/ICCV.2015.286</identifier><identifier>CODEN: IEEPAD</identifier><language>eng</language><publisher>IEEE</publisher><subject>Computational modeling ; Computer vision ; Conferences ; History ; Image detection ; Learning ; Learning (artificial intelligence) ; Localization ; Position (location) ; Prediction algorithms ; Proposals ; Reinforcement ; Search problems ; Transformations ; Transforms</subject><ispartof>2015 IEEE International Conference on Computer Vision (ICCV), 2015, p.2488-2496</ispartof><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/7410643$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,314,777,781,786,787,23911,23912,25121,27905,27906,54536,54913</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/7410643$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Caicedo, Juan C.</creatorcontrib><creatorcontrib>Lazebnik, Svetlana</creatorcontrib><title>Active Object Localization with Deep Reinforcement Learning</title><title>2015 IEEE International Conference on Computer Vision (ICCV)</title><addtitle>ICCV</addtitle><description>We present an active detection model for localizing objects in scenes. The model is class-specific and allows an agent to focus attention on candidate regions for identifying the correct location of a target object. This agent learns to deform a bounding box using simple transformation actions, with the goal of determining the most specific location of target objects following top-down reasoning. The proposed localization agent is trained using deep reinforcement learning, and evaluated on the Pascal VOC 2007 dataset. We show that agents guided by the proposed model are able to localize a single instance of an object after analyzing only between 11 and 25 regions in an image, and obtain the best detection results among systems that do not use object proposals for object localization.</description><subject>Computational modeling</subject><subject>Computer vision</subject><subject>Conferences</subject><subject>History</subject><subject>Image detection</subject><subject>Learning</subject><subject>Learning (artificial intelligence)</subject><subject>Localization</subject><subject>Position (location)</subject><subject>Prediction algorithms</subject><subject>Proposals</subject><subject>Reinforcement</subject><subject>Search problems</subject><subject>Transformations</subject><subject>Transforms</subject><issn>2380-7504</issn><isbn>1467383910</isbn><isbn>9781467383912</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2015</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNotzLFLw0AUgPFTEGyro5NLRpfU9_LuLnc4lVi1ECiIuobL5UWvpElNUkX_egt1-pYfnxBXCHNEsLerLHubJ4Bqnhh9IqYodUqGLMKpmCRkIE4VyHMxHYYNANmDmoi7hR_DF0frcsN-jPLOuyb8ujF0bfQdxo_onnkXPXNo6673vOX2gNj1bWjfL8RZ7ZqBL_87E68Py5fsKc7Xj6tskcchkekYk6pQAUqlarYl-DoF7SsGMCU6adk5XbmqVOQNU1lBYgCUUSkp1gkoSzNxc_zu-u5zz8NYbMPguWlcy91-KNCgBm2R6ECvjzQwc7Hrw9b1P0UqEbQk-gOjtFN5</recordid><startdate>20151201</startdate><enddate>20151201</enddate><creator>Caicedo, Juan C.</creator><creator>Lazebnik, Svetlana</creator><general>IEEE</general><scope>6IE</scope><scope>6IH</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIO</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope></search><sort><creationdate>20151201</creationdate><title>Active Object Localization with Deep Reinforcement Learning</title><author>Caicedo, Juan C. ; Lazebnik, Svetlana</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i247t-35d1501455fe9b0cf706cde008b1a49eaa6dadb53c8e3bd02800585735e620593</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2015</creationdate><topic>Computational modeling</topic><topic>Computer vision</topic><topic>Conferences</topic><topic>History</topic><topic>Image detection</topic><topic>Learning</topic><topic>Learning (artificial intelligence)</topic><topic>Localization</topic><topic>Position (location)</topic><topic>Prediction algorithms</topic><topic>Proposals</topic><topic>Reinforcement</topic><topic>Search problems</topic><topic>Transformations</topic><topic>Transforms</topic><toplevel>online_resources</toplevel><creatorcontrib>Caicedo, Juan C.</creatorcontrib><creatorcontrib>Lazebnik, Svetlana</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan (POP) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE/IET Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP) 1998-present</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Caicedo, Juan C.</au><au>Lazebnik, Svetlana</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Active Object Localization with Deep Reinforcement Learning</atitle><btitle>2015 IEEE International Conference on Computer Vision (ICCV)</btitle><stitle>ICCV</stitle><date>2015-12-01</date><risdate>2015</risdate><spage>2488</spage><epage>2496</epage><pages>2488-2496</pages><eissn>2380-7504</eissn><eisbn>1467383910</eisbn><eisbn>9781467383912</eisbn><coden>IEEPAD</coden><abstract>We present an active detection model for localizing objects in scenes. The model is class-specific and allows an agent to focus attention on candidate regions for identifying the correct location of a target object. This agent learns to deform a bounding box using simple transformation actions, with the goal of determining the most specific location of target objects following top-down reasoning. The proposed localization agent is trained using deep reinforcement learning, and evaluated on the Pascal VOC 2007 dataset. We show that agents guided by the proposed model are able to localize a single instance of an object after analyzing only between 11 and 25 regions in an image, and obtain the best detection results among systems that do not use object proposals for object localization.</abstract><pub>IEEE</pub><doi>10.1109/ICCV.2015.286</doi><tpages>9</tpages><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier EISSN: 2380-7504
ispartof 2015 IEEE International Conference on Computer Vision (ICCV), 2015, p.2488-2496
issn 2380-7504
language eng
recordid cdi_proquest_miscellaneous_1816069133
source IEEE Xplore All Conference Series
subjects Computational modeling
Computer vision
Conferences
History
Image detection
Learning
Learning (artificial intelligence)
Localization
Position (location)
Prediction algorithms
Proposals
Reinforcement
Search problems
Transformations
Transforms
title Active Object Localization with Deep Reinforcement Learning
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-19T23%3A48%3A26IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Active%20Object%20Localization%20with%20Deep%20Reinforcement%20Learning&rft.btitle=2015%20IEEE%20International%20Conference%20on%20Computer%20Vision%20(ICCV)&rft.au=Caicedo,%20Juan%20C.&rft.date=2015-12-01&rft.spage=2488&rft.epage=2496&rft.pages=2488-2496&rft.eissn=2380-7504&rft.coden=IEEPAD&rft_id=info:doi/10.1109/ICCV.2015.286&rft.eisbn=1467383910&rft.eisbn_list=9781467383912&rft_dat=%3Cproquest_CHZPO%3E1816069133%3C/proquest_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i247t-35d1501455fe9b0cf706cde008b1a49eaa6dadb53c8e3bd02800585735e620593%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=1816069133&rft_id=info:pmid/&rft_ieee_id=7410643&rfr_iscdi=true