Multiscale Spectral Residue for Faster Image Object Detection

Bibliographic Details
Main Author: Silva Filho, Jose Grimaldo da
Publication Date: 2013
Format: Master thesis
Language: por
Source: Repositório Institucional da UFBA
Download full: http://repositorio.ufba.br/ri/handle/ri/21340
Summary: Accuracy in image object detection has been usually achieved at the expense of much computational load. Therefore a trade-o between detection performance and fast execution commonly represents the ultimate goal of an object detector in real life applications. Most images are composed of non-trivial amounts of background information, such as sky, ground and water. In this sense, using an object detector against a recurring background pattern can require a signi cant amount of the total processing time. To alleviate this problem, search space reduction methods can help focusing the detection procedure on more distinctive image regions.
id UFBA-2_a1c3767a6c0fcf1d839f8c9305d4b17d
oai_identifier_str oai:repositorio.ufba.br:ri/21340
network_acronym_str UFBA-2
network_name_str Repositório Institucional da UFBA
repository_id_str 1932
spelling Silva Filho, Jose Grimaldo daOliveira, Luciano Rebouças deSchwartz, William RobsonMello, Vinicius MoreiraApolinário Júnior, Antonio Lopes2017-02-07T11:51:58Z2017-02-07T11:51:58Z2017-02-072013-01-18http://repositorio.ufba.br/ri/handle/ri/21340Accuracy in image object detection has been usually achieved at the expense of much computational load. Therefore a trade-o between detection performance and fast execution commonly represents the ultimate goal of an object detector in real life applications. Most images are composed of non-trivial amounts of background information, such as sky, ground and water. In this sense, using an object detector against a recurring background pattern can require a signi cant amount of the total processing time. To alleviate this problem, search space reduction methods can help focusing the detection procedure on more distinctive image regions.Among the several approaches for search space reduction, we explored saliency information to organize regions based on their probability of containing objects. Saliency detectors are capable of pinpointing regions which generate stronger visual stimuli based solely on information extracted from the image. The fact that saliency methods do not require prior training is an important benefit, which allows application of these techniques in a broad range of machine vision domains. We propose a novel method toward the goal of faster object detectors. The proposed method was grounded on a multi-scale spectral residue (MSR) analysis using saliency detection. For better search space reduction, our method enables fine control of search scale, more robustness to variations on saliency intensity along an object length and also a direct way to control the balance between search space reduction and false negatives caused by region selection. Compared to a regular sliding window search over the images, in our experiments, MSR was able to reduce by 75% (in average) the number of windows to be evaluated by an object detector while improving or at least maintaining detector ROC performance. The proposed method was thoroughly evaluated over a subset of LabelMe dataset (person images), improving detection performance in most cases. This evaluation was done comparing object detection performance against different object detectors, with and without MSR. Additionally, we also provide evaluation of how different object classes interact with MSR, which was done using Pascal VOC 2007 dataset. Finally, tests made showed that window selection performance of MSR has a good scalability with regard to image size. From the obtained data, our conclusion is that MSR can provide substantial benefits to existing sliding window detectorsSubmitted by Diogo Barreiros (diogo.barreiros@ufba.br) on 2017-02-06T16:59:36Z No. of bitstreams: 1 dissertacao_mestrado_jose-grimaldo.pdf: 19406681 bytes, checksum: d108855fa0fb0d44ee5d1cb59579a04c (MD5)Approved for entry into archive by Vanessa Reis (vanessa.jamile@ufba.br) on 2017-02-07T11:51:58Z (GMT) No. of bitstreams: 1 dissertacao_mestrado_jose-grimaldo.pdf: 19406681 bytes, checksum: d108855fa0fb0d44ee5d1cb59579a04c (MD5)Made available in DSpace on 2017-02-07T11:51:58Z (GMT). No. of bitstreams: 1 dissertacao_mestrado_jose-grimaldo.pdf: 19406681 bytes, checksum: d108855fa0fb0d44ee5d1cb59579a04c (MD5)Multiscale Spectral ResidueImageObjectMultiscale Spectral Residue for Faster Image Object Detectioninfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisEscola Politécnica / Instituto de MatemáticaPrograma de Pós-Graduação em MecatrônicaUFBAbrasilinfo:eu-repo/semantics/openAccessporreponame:Repositório Institucional da UFBAinstname:Universidade Federal da Bahia (UFBA)instacron:UFBAORIGINALdissertacao_mestrado_jose-grimaldo.pdfdissertacao_mestrado_jose-grimaldo.pdfapplication/pdf19406681https://repositorio.ufba.br/bitstream/ri/21340/1/dissertacao_mestrado_jose-grimaldo.pdfd108855fa0fb0d44ee5d1cb59579a04cMD51LICENSElicense.txtlicense.txttext/plain1345https://repositorio.ufba.br/bitstream/ri/21340/2/license.txtff6eaa8b858ea317fded99f125f5fcd0MD52TEXTdissertacao_mestrado_jose-grimaldo.pdf.txtdissertacao_mestrado_jose-grimaldo.pdf.txtExtracted texttext/plain208826https://repositorio.ufba.br/bitstream/ri/21340/3/dissertacao_mestrado_jose-grimaldo.pdf.txte0d792610b0ba093db14db46d161ebaaMD53ri/213402022-07-01 10:46:41.901oai:repositorio.ufba.br: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Repositório InstitucionalPUBhttps://repositorio.ufba.br/oai/requestrepositorio@ufba.bropendoar:19322022-07-01T13:46:41Repositório Institucional da UFBA - Universidade Federal da Bahia (UFBA)false
dc.title.pt_BR.fl_str_mv Multiscale Spectral Residue for Faster Image Object Detection
title Multiscale Spectral Residue for Faster Image Object Detection
spellingShingle Multiscale Spectral Residue for Faster Image Object Detection
Silva Filho, Jose Grimaldo da
Multiscale Spectral Residue
Image
Object
title_short Multiscale Spectral Residue for Faster Image Object Detection
title_full Multiscale Spectral Residue for Faster Image Object Detection
title_fullStr Multiscale Spectral Residue for Faster Image Object Detection
title_full_unstemmed Multiscale Spectral Residue for Faster Image Object Detection
title_sort Multiscale Spectral Residue for Faster Image Object Detection
author Silva Filho, Jose Grimaldo da
author_facet Silva Filho, Jose Grimaldo da
author_role author
dc.contributor.author.fl_str_mv Silva Filho, Jose Grimaldo da
dc.contributor.advisor1.fl_str_mv Oliveira, Luciano Rebouças de
dc.contributor.referee1.fl_str_mv Schwartz, William Robson
Mello, Vinicius Moreira
Apolinário Júnior, Antonio Lopes
contributor_str_mv Oliveira, Luciano Rebouças de
Schwartz, William Robson
Mello, Vinicius Moreira
Apolinário Júnior, Antonio Lopes
dc.subject.por.fl_str_mv Multiscale Spectral Residue
Image
Object
topic Multiscale Spectral Residue
Image
Object
description Accuracy in image object detection has been usually achieved at the expense of much computational load. Therefore a trade-o between detection performance and fast execution commonly represents the ultimate goal of an object detector in real life applications. Most images are composed of non-trivial amounts of background information, such as sky, ground and water. In this sense, using an object detector against a recurring background pattern can require a signi cant amount of the total processing time. To alleviate this problem, search space reduction methods can help focusing the detection procedure on more distinctive image regions.
publishDate 2013
dc.date.submitted.none.fl_str_mv 2013-01-18
dc.date.accessioned.fl_str_mv 2017-02-07T11:51:58Z
dc.date.available.fl_str_mv 2017-02-07T11:51:58Z
dc.date.issued.fl_str_mv 2017-02-07
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://repositorio.ufba.br/ri/handle/ri/21340
url http://repositorio.ufba.br/ri/handle/ri/21340
dc.language.iso.fl_str_mv por
language por
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.publisher.none.fl_str_mv Escola Politécnica / Instituto de Matemática
dc.publisher.program.fl_str_mv Programa de Pós-Graduação em Mecatrônica
dc.publisher.initials.fl_str_mv UFBA
dc.publisher.country.fl_str_mv brasil
publisher.none.fl_str_mv Escola Politécnica / Instituto de Matemática
dc.source.none.fl_str_mv reponame:Repositório Institucional da UFBA
instname:Universidade Federal da Bahia (UFBA)
instacron:UFBA
instname_str Universidade Federal da Bahia (UFBA)
instacron_str UFBA
institution UFBA
reponame_str Repositório Institucional da UFBA
collection Repositório Institucional da UFBA
bitstream.url.fl_str_mv https://repositorio.ufba.br/bitstream/ri/21340/1/dissertacao_mestrado_jose-grimaldo.pdf
https://repositorio.ufba.br/bitstream/ri/21340/2/license.txt
https://repositorio.ufba.br/bitstream/ri/21340/3/dissertacao_mestrado_jose-grimaldo.pdf.txt
bitstream.checksum.fl_str_mv d108855fa0fb0d44ee5d1cb59579a04c
ff6eaa8b858ea317fded99f125f5fcd0
e0d792610b0ba093db14db46d161ebaa
bitstream.checksumAlgorithm.fl_str_mv MD5
MD5
MD5
repository.name.fl_str_mv Repositório Institucional da UFBA - Universidade Federal da Bahia (UFBA)
repository.mail.fl_str_mv repositorio@ufba.br
_version_ 1847339199572738048

Similar Items