Close

1. Identity statement
Reference TypeConference Paper (Conference Proceedings)
Sitesibgrapi.sid.inpe.br
Holder Codeibi 8JMKD3MGPEW34M/46T9EHH
Identifier8JMKD3MGPEW34M/45C6UK8
Repositorysid.inpe.br/sibgrapi/2021/09.01.22.31
Last Update2021:09.02.22.41.31 (UTC) administrator
Metadata Repositorysid.inpe.br/sibgrapi/2021/09.01.22.31.22
Metadata Last Update2022:06.14.00.00.19 (UTC) administrator
DOI10.1109/SIBGRAPI54419.2021.00035
Citation KeyBlangerHiraJian:2021:ReNeBo
TitleReducing the need for bounding box annotations in Object Detection using Image Classification data
FormatOn-line
Year2021
Access Date2022, Dec. 07
Number of Files1
Size4045 KiB
2. Context
Author1 Blanger, Leonardo
2 Hirata, Nina S. T.
3 Jiang, Xiaoyi
Affiliation1 University of Sao Paulo 
2 University of Sao Paulo 
3 University of Münster
EditorPaiva, Afonso
Menotti, David
Baranoski, Gladimir V. G.
Proença, Hugo Pedro
Junior, Antonio Lopes Apolinario
Papa, João Paulo
Pagliosa, Paulo
dos Santos, Thiago Oliveira
e Sá, Asla Medeiros
da Silveira, Thiago Lopes Trugillo
Brazil, Emilio Vital
Ponti, Moacir A.
Fernandes, Leandro A. F.
Avila, Sandra
e-Mail Addressleonardoblanger@gmail.com
Conference NameConference on Graphics, Patterns and Images, 34 (SIBGRAPI)
Conference LocationGramado, RS, Brazil (virtual)
Date18-22 Oct. 2021
PublisherIEEE Computer Society
Publisher CityLos Alamitos
Book TitleProceedings
Tertiary TypeFull Paper
History (UTC)2021-09-02 22:41:31 :: leonardoblanger@gmail.com -> administrator :: 2021
2022-03-02 00:54:15 :: administrator -> menottid@gmail.com :: 2021
2022-03-02 13:21:38 :: menottid@gmail.com -> administrator :: 2021
2022-06-14 00:00:19 :: administrator -> :: 2021
3. Content and structure
Is the master or a copy?is the master
Content Stagecompleted
Transferable1
Version Typefinaldraft
Keywordssample synthesis
object detection
pretraining
deep learning
AbstractWe address the problem of training Object Detection models using significantly less bounding box annotated images. For that, we take advantage of cheaper and more abundant image classification data. Our proposal consists in automatically generating artificial detection samples, with no need of expensive detection level supervision, using images with classification labels only. We also detail a pretraining initialization strategy for detection architectures using these artificially synthesized samples, before finetuning on real detection data, and experimentally show how this consistently leads to more data efficient models. With the proposed approach, we were able to effectively use only classification data to improve results on the harder and more supervision hungry object detection problem. We achieve results equivalent to those of the full data scenario using only a small fraction of the original detection data for Face, Bird, and Car detection.
Arrangement 1urlib.net > SDLA > Fonds > SIBGRAPI 2021 > Reducing the need...
Arrangement 2urlib.net > SDLA > Fonds > Full Index > Reducing the need...
doc Directory Contentaccess
source Directory Contentthere are no files
agreement Directory Content
agreement.html 01/09/2021 19:31 1.3 KiB 
4. Conditions of access and use
data URLhttp://urlib.net/ibi/8JMKD3MGPEW34M/45C6UK8
zipped data URLhttp://urlib.net/zip/8JMKD3MGPEW34M/45C6UK8
Languageen
Target File29.pdf
User Groupleonardoblanger@gmail.com
Visibilityshown
Update Permissionnot transferred
5. Allied materials
Mirror Repositorysid.inpe.br/banon/2001/03.30.15.38.24
Next Higher Units8JMKD3MGPEW34M/45PQ3RS
8JMKD3MGPEW34M/4742MCS
Host Collectionsid.inpe.br/banon/2001/03.30.15.38
6. Notes
Empty Fieldsarchivingpolicy archivist area callnumber contenttype copyholder copyright creatorhistory descriptionlevel dissemination edition electronicmailaddress group isbn issn label lineage mark nextedition notes numberofvolumes orcid organization pages parameterlist parentrepositories previousedition previouslowerunit progress project readergroup readpermission resumeid rightsholder secondarydate secondarykey secondarymark secondarytype serieseditor session shorttitle sponsor subject tertiarymark type url volume


Close