Identity statement area | |
Reference Type | Conference Paper (Conference Proceedings) |
Site | sibgrapi.sid.inpe.br |
Identifier | 8JMKD3MGPAW/3PHKBHE |
Repository | sid.inpe.br/sibgrapi/2017/09.01.19.59 |
Last Update | 2017:09.01.19.59.43 administrator |
Metadata | sid.inpe.br/sibgrapi/2017/09.01.19.59.43 |
Metadata Last Update | 2020:02.19.02.01.43 administrator |
Citation Key | AlmeidaJrGuim:2017:CaStHu |
Title | A New Pooling Strategy based on Local Feature Distribution: A Case Study for Human Action Classification  |
Format | On-line |
Year | 2017 |
Date | Oct. 17-20, 2017 |
Access Date | 2021, Jan. 21 |
Number of Files | 1 |
Size | 2501 KiB |
Context area | |
Author | 1 Almeida, Raquel 2 Jr, Zenilton K. Goncalves do Patrocinio 3 Guimaraes, Silvio Jamil Ferzoli |
Affiliation | 1 PUC Minas 2 PUC Minas 3 PUC Minas |
Editor | Torchelsen, Rafael Piccin Nascimento, Erickson Rangel do Panozzo, Daniele Liu, Zicheng Farias, Mylène Viera, Thales Sacht, Leonardo Ferreira, Nivan Comba, João Luiz Dihl Hirata, Nina Schiavon Porto, Marcelo Vital, Creto Pagot, Christian Azambuja Petronetto, Fabiano Clua, Esteban Cardeal, Flávio |
e-Mail Address | sjamil@pucminas.br |
Conference Name | Conference on Graphics, Patterns and Images, 30 (SIBGRAPI) |
Conference Location | Niterói, RJ |
Book Title | Proceedings |
Publisher | IEEE Computer Society |
Publisher City | Los Alamitos |
Tertiary Type | Full Paper |
History | 2017-09-01 19:59:43 :: sjamil@pucminas.br -> administrator :: 2020-02-19 02:01:43 :: administrator -> :: 2017 |
Content and structure area | |
Is the master or a copy? | is the master |
Content Stage | completed |
Transferable | 1 |
Content Type | External Contribution |
Keywords | Mid-level representation, Human action classification, Pooling. |
Abstract | Mid-level representations are used to map sets of local features into one global representation for a given media descriptor. In visual pattern recognition tasks, Bag-of-Words (BoW) is one popular strategy, among many methods available in literature, due mainly by the simplicity in concept and implementation. Despite the overall good results achieved by BoW in many tasks, the method is unstable in high dimensional feature space and quantization errors are usually ignored in the final representation. To cope with these problems, we propose a new pooling function based on feature points distribution around codewords. We propose to use the standard deviation associated with each codeword to measure attribution discrepancy and weight the impact that feature points will assume in the final representation. The main contribution of this article is the study of more discriminative representations, which amplify values of feature points close to codewords border regions. Experiments were conducted in human action classification task and results demonstrated that our pooling strategy has improved the classification rates in 25.6% for UCF Sports dataset and 21.4% for UCF 11 dataset, with respect to the original pooling function used in BoW. |
source Directory Content | there are no files |
agreement Directory Content | |
Conditions of access and use area | |
data URL | http://urlib.net/rep/8JMKD3MGPAW/3PHKBHE |
zipped data URL | http://urlib.net/zip/8JMKD3MGPAW/3PHKBHE |
Language | en |
Target File | PID4982925.pdf |
User Group | sjamil@pucminas.br |
Visibility | shown |
Update Permission | not transferred |
Allied materials area | |
Mirror Repository | sid.inpe.br/banon/2001/03.30.15.38.24 |
Next Higher Units | 8JMKD3MGPAW/3PJT9LS 8JMKD3MGPAW/3PKCC58 |
Host Collection | sid.inpe.br/banon/2001/03.30.15.38 |
Notes area | |
Empty Fields | accessionnumber archivingpolicy archivist area callnumber copyholder copyright creatorhistory descriptionlevel dissemination doi edition electronicmailaddress group holdercode 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 versiontype volume |
| |