Learning ordered pooling weights in image classification

Spatial pooling is an important step in computer vision systems like Convolutional Neural Networks or the Bag-of-Words method. The spatial pooling purpose is to combine neighbouring descriptors to obtain a single descriptor for a given region (local or global). The resultant combined vector must be...

Descripción completa

Detalles Bibliográficos
Autores: Forcén Carvalho, Juan Ignacio, Pagola Barrio, Miguel, Barrenechea Tartas, Edurne, Bustince Sola, Humberto
Tipo de recurso: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2020
País:España
Institución:Universidad Pública de Navarra
Repositorio:Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
OAI Identifier:oai:academica-e.unavarra.es:2454/39499
Acceso en línea:https://hdl.handle.net/2454/39499
Access Level:acceso abierto
Palabra clave:Pooling
Ordered weighted aggregation
Image classification
Bag-of-words
Mid-level features
Convolutional neural networks
Global pooling
id ES_2bfe23556fc40bee4f89bc51bf8d3f67
oai_identifier_str oai:academica-e.unavarra.es:2454/39499
network_acronym_str ES
network_name_str España
repository_id_str
spelling Learning ordered pooling weights in image classificationForcén Carvalho, Juan IgnacioPagola Barrio, MiguelBarrenechea Tartas, EdurneBustince Sola, HumbertoPoolingOrdered weighted aggregationImage classificationBag-of-wordsMid-level featuresConvolutional neural networksGlobal poolingSpatial pooling is an important step in computer vision systems like Convolutional Neural Networks or the Bag-of-Words method. The spatial pooling purpose is to combine neighbouring descriptors to obtain a single descriptor for a given region (local or global). The resultant combined vector must be as discriminant as possible, in other words, must contain relevant information, while removing irrelevant and confusing details. Maximum and average are the most common aggregation functions used in the pooling step. To improve the aggregation of relevant information without degrading their discriminative power for image classification, we introduce a simple but effective scheme based on Ordered Weighted Average (OWA) aggregation operators. We present a method to learn the weights of the OWA aggregation operator in a Bag-of-Words framework and in Convolutional Neural Networks, and provide an extensive evaluation showing that OWA based pooling outperforms classical aggregation operators.This work is partially supported by the research services of Universidad Pública de Navarra and by the project TIN2016-77356-P (AEI/FEDER, UE).ElsevierEstadística, Informática y MatemáticasEstatistika, Informatika eta MatematikaUniversidad Pública de Navarra / Nafarroako Unibertsitate Publikoa2020info:eu-repo/semantics/articleinfo:eu-repo/semantics/acceptedVersionapplication/pdfhttps://hdl.handle.net/2454/39499reponame:Academica-e. Repositorio Institucional de la Universidad Pública de Navarrainstname:Universidad Pública de NavarraInglésinfo:eu-repo/grantAgreement/ES/1PE/TIN2016-77356-P© 2020 Elsevier B.V. This manuscript version is made available under the CC-BY-NC-ND 4.1https://creativecommons.org/licenses/by-nc-nd/4.0/info:eu-repo/semantics/openAccessoai:academica-e.unavarra.es:2454/394992026-06-17T12:41:47Z
dc.title.none.fl_str_mv Learning ordered pooling weights in image classification
title Learning ordered pooling weights in image classification
spellingShingle Learning ordered pooling weights in image classification
Forcén Carvalho, Juan Ignacio
Pooling
Ordered weighted aggregation
Image classification
Bag-of-words
Mid-level features
Convolutional neural networks
Global pooling
title_short Learning ordered pooling weights in image classification
title_full Learning ordered pooling weights in image classification
title_fullStr Learning ordered pooling weights in image classification
title_full_unstemmed Learning ordered pooling weights in image classification
title_sort Learning ordered pooling weights in image classification
dc.creator.none.fl_str_mv Forcén Carvalho, Juan Ignacio
Pagola Barrio, Miguel
Barrenechea Tartas, Edurne
Bustince Sola, Humberto
author Forcén Carvalho, Juan Ignacio
author_facet Forcén Carvalho, Juan Ignacio
Pagola Barrio, Miguel
Barrenechea Tartas, Edurne
Bustince Sola, Humberto
author_role author
author2 Pagola Barrio, Miguel
Barrenechea Tartas, Edurne
Bustince Sola, Humberto
author2_role author
author
author
dc.contributor.none.fl_str_mv Estadística, Informática y Matemáticas
Estatistika, Informatika eta Matematika
Universidad Pública de Navarra / Nafarroako Unibertsitate Publikoa
dc.subject.none.fl_str_mv Pooling
Ordered weighted aggregation
Image classification
Bag-of-words
Mid-level features
Convolutional neural networks
Global pooling
topic Pooling
Ordered weighted aggregation
Image classification
Bag-of-words
Mid-level features
Convolutional neural networks
Global pooling
description Spatial pooling is an important step in computer vision systems like Convolutional Neural Networks or the Bag-of-Words method. The spatial pooling purpose is to combine neighbouring descriptors to obtain a single descriptor for a given region (local or global). The resultant combined vector must be as discriminant as possible, in other words, must contain relevant information, while removing irrelevant and confusing details. Maximum and average are the most common aggregation functions used in the pooling step. To improve the aggregation of relevant information without degrading their discriminative power for image classification, we introduce a simple but effective scheme based on Ordered Weighted Average (OWA) aggregation operators. We present a method to learn the weights of the OWA aggregation operator in a Bag-of-Words framework and in Convolutional Neural Networks, and provide an extensive evaluation showing that OWA based pooling outperforms classical aggregation operators.
publishDate 2020
dc.date.none.fl_str_mv 2020
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/acceptedVersion
format article
status_str acceptedVersion
dc.identifier.none.fl_str_mv https://hdl.handle.net/2454/39499
url https://hdl.handle.net/2454/39499
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv info:eu-repo/grantAgreement/ES/1PE/TIN2016-77356-P
dc.rights.none.fl_str_mv © 2020 Elsevier B.V. This manuscript version is made available under the CC-BY-NC-ND 4.1
https://creativecommons.org/licenses/by-nc-nd/4.0/
info:eu-repo/semantics/openAccess
rights_invalid_str_mv © 2020 Elsevier B.V. This manuscript version is made available under the CC-BY-NC-ND 4.1
https://creativecommons.org/licenses/by-nc-nd/4.0/
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Elsevier
publisher.none.fl_str_mv Elsevier
dc.source.none.fl_str_mv reponame:Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
instname:Universidad Pública de Navarra
instname_str Universidad Pública de Navarra
reponame_str Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
collection Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869405198616625152
score 15,81155

Similares en LA Referencia