Loading…

MS-RMAC: Multiscale Regional Maximum Activation of Convolutions for Image Retrieval

Recent works have demonstrated that image descriptors produced by convolutional feature maps provide state-of-the-art performance for image retrieval and classification problems. However, features from a single convolutional layer are not robust enough for shape deformation, scale variation, and hea...

Full description

Saved in:
Bibliographic Details
Published in:IEEE signal processing letters 2017-05, Vol.24 (5), p.609-613
Main Authors: Li, Yang, Xu, Yulong, Wang, Jiabao, Miao, Zhuang, Zhang, Yafei
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3
cites cdi_FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3
container_end_page 613
container_issue 5
container_start_page 609
container_title IEEE signal processing letters
container_volume 24
creator Li, Yang
Xu, Yulong
Wang, Jiabao
Miao, Zhuang
Zhang, Yafei
description Recent works have demonstrated that image descriptors produced by convolutional feature maps provide state-of-the-art performance for image retrieval and classification problems. However, features from a single convolutional layer are not robust enough for shape deformation, scale variation, and heavy occlusion. In this letter, we present a simple and straightforward approach for extracting multiscale (MS) regional maximum activation of convolutions features from different layers of the convolutional neural network. And we also propose aggregating MS features into a single vector by a parameter-free hedge method for image retrieval. Extensive experimental results on three challenging benchmark datasets indicate that the proposed method achieved outstanding performance against state-of-the-art methods.
doi_str_mv 10.1109/LSP.2017.2665522
format article
fullrecord <record><control><sourceid>crossref_ieee_</sourceid><recordid>TN_cdi_ieee_primary_7845566</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>7845566</ieee_id><sourcerecordid>10_1109_LSP_2017_2665522</sourcerecordid><originalsourceid>FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3</originalsourceid><addsrcrecordid>eNo9kFFLwzAQx4MoOKfvgi_5Ap2XpEkT30bROVhRtr2XtL2MSLtK0w799rZs-HR3f-53HD9CHhksGAPzvNl9LjiwZMGVkpLzKzJjUuqIC8Wuxx4SiIwBfUvuQvgCAM20nJFdtou22TJ9odlQ9z6Utka6xYNvj7ammf3xzdDQZdn7k-3HkLaOpu3x1NbDNAbq2o6uG3uYqL7zeLL1Pblxtg74cKlzsn973afv0eZjtU6Xm6jkSvSRVkooDVIIp8vYMSHQaRabWKBKjKxAozWyqJgCpXVVSmsKKQqlkCfIrJgTOJ8tuzaEDl3-3fnGdr85g3xyko9O8slJfnEyIk9nxCPi_3qiYynHZ_4AMT9cnw</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>MS-RMAC: Multiscale Regional Maximum Activation of Convolutions for Image Retrieval</title><source>IEEE Electronic Library (IEL) Journals</source><creator>Li, Yang ; Xu, Yulong ; Wang, Jiabao ; Miao, Zhuang ; Zhang, Yafei</creator><creatorcontrib>Li, Yang ; Xu, Yulong ; Wang, Jiabao ; Miao, Zhuang ; Zhang, Yafei</creatorcontrib><description>Recent works have demonstrated that image descriptors produced by convolutional feature maps provide state-of-the-art performance for image retrieval and classification problems. However, features from a single convolutional layer are not robust enough for shape deformation, scale variation, and heavy occlusion. In this letter, we present a simple and straightforward approach for extracting multiscale (MS) regional maximum activation of convolutions features from different layers of the convolutional neural network. And we also propose aggregating MS features into a single vector by a parameter-free hedge method for image retrieval. Extensive experimental results on three challenging benchmark datasets indicate that the proposed method achieved outstanding performance against state-of-the-art methods.</description><identifier>ISSN: 1070-9908</identifier><identifier>EISSN: 1558-2361</identifier><identifier>DOI: 10.1109/LSP.2017.2665522</identifier><identifier>CODEN: ISPLEM</identifier><language>eng</language><publisher>IEEE</publisher><subject>Convolution ; Convolutional codes ; Convolutional neural network (CNN) ; Feature extraction ; Image retrieval ; multiscale (MS) feature ; Neural networks ; Robustness ; Signal processing algorithms</subject><ispartof>IEEE signal processing letters, 2017-05, Vol.24 (5), p.609-613</ispartof><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3</citedby><cites>FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/7845566$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,27923,27924,54795</link.rule.ids></links><search><creatorcontrib>Li, Yang</creatorcontrib><creatorcontrib>Xu, Yulong</creatorcontrib><creatorcontrib>Wang, Jiabao</creatorcontrib><creatorcontrib>Miao, Zhuang</creatorcontrib><creatorcontrib>Zhang, Yafei</creatorcontrib><title>MS-RMAC: Multiscale Regional Maximum Activation of Convolutions for Image Retrieval</title><title>IEEE signal processing letters</title><addtitle>LSP</addtitle><description>Recent works have demonstrated that image descriptors produced by convolutional feature maps provide state-of-the-art performance for image retrieval and classification problems. However, features from a single convolutional layer are not robust enough for shape deformation, scale variation, and heavy occlusion. In this letter, we present a simple and straightforward approach for extracting multiscale (MS) regional maximum activation of convolutions features from different layers of the convolutional neural network. And we also propose aggregating MS features into a single vector by a parameter-free hedge method for image retrieval. Extensive experimental results on three challenging benchmark datasets indicate that the proposed method achieved outstanding performance against state-of-the-art methods.</description><subject>Convolution</subject><subject>Convolutional codes</subject><subject>Convolutional neural network (CNN)</subject><subject>Feature extraction</subject><subject>Image retrieval</subject><subject>multiscale (MS) feature</subject><subject>Neural networks</subject><subject>Robustness</subject><subject>Signal processing algorithms</subject><issn>1070-9908</issn><issn>1558-2361</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2017</creationdate><recordtype>article</recordtype><recordid>eNo9kFFLwzAQx4MoOKfvgi_5Ap2XpEkT30bROVhRtr2XtL2MSLtK0w799rZs-HR3f-53HD9CHhksGAPzvNl9LjiwZMGVkpLzKzJjUuqIC8Wuxx4SiIwBfUvuQvgCAM20nJFdtou22TJ9odlQ9z6Utka6xYNvj7ammf3xzdDQZdn7k-3HkLaOpu3x1NbDNAbq2o6uG3uYqL7zeLL1Pblxtg74cKlzsn973afv0eZjtU6Xm6jkSvSRVkooDVIIp8vYMSHQaRabWKBKjKxAozWyqJgCpXVVSmsKKQqlkCfIrJgTOJ8tuzaEDl3-3fnGdr85g3xyko9O8slJfnEyIk9nxCPi_3qiYynHZ_4AMT9cnw</recordid><startdate>201705</startdate><enddate>201705</enddate><creator>Li, Yang</creator><creator>Xu, Yulong</creator><creator>Wang, Jiabao</creator><creator>Miao, Zhuang</creator><creator>Zhang, Yafei</creator><general>IEEE</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>201705</creationdate><title>MS-RMAC: Multiscale Regional Maximum Activation of Convolutions for Image Retrieval</title><author>Li, Yang ; Xu, Yulong ; Wang, Jiabao ; Miao, Zhuang ; Zhang, Yafei</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2017</creationdate><topic>Convolution</topic><topic>Convolutional codes</topic><topic>Convolutional neural network (CNN)</topic><topic>Feature extraction</topic><topic>Image retrieval</topic><topic>multiscale (MS) feature</topic><topic>Neural networks</topic><topic>Robustness</topic><topic>Signal processing algorithms</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Li, Yang</creatorcontrib><creatorcontrib>Xu, Yulong</creatorcontrib><creatorcontrib>Wang, Jiabao</creatorcontrib><creatorcontrib>Miao, Zhuang</creatorcontrib><creatorcontrib>Zhang, Yafei</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE/IET Electronic Library</collection><collection>CrossRef</collection><jtitle>IEEE signal processing letters</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Li, Yang</au><au>Xu, Yulong</au><au>Wang, Jiabao</au><au>Miao, Zhuang</au><au>Zhang, Yafei</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>MS-RMAC: Multiscale Regional Maximum Activation of Convolutions for Image Retrieval</atitle><jtitle>IEEE signal processing letters</jtitle><stitle>LSP</stitle><date>2017-05</date><risdate>2017</risdate><volume>24</volume><issue>5</issue><spage>609</spage><epage>613</epage><pages>609-613</pages><issn>1070-9908</issn><eissn>1558-2361</eissn><coden>ISPLEM</coden><abstract>Recent works have demonstrated that image descriptors produced by convolutional feature maps provide state-of-the-art performance for image retrieval and classification problems. However, features from a single convolutional layer are not robust enough for shape deformation, scale variation, and heavy occlusion. In this letter, we present a simple and straightforward approach for extracting multiscale (MS) regional maximum activation of convolutions features from different layers of the convolutional neural network. And we also propose aggregating MS features into a single vector by a parameter-free hedge method for image retrieval. Extensive experimental results on three challenging benchmark datasets indicate that the proposed method achieved outstanding performance against state-of-the-art methods.</abstract><pub>IEEE</pub><doi>10.1109/LSP.2017.2665522</doi><tpages>5</tpages></addata></record>
fulltext fulltext
identifier ISSN: 1070-9908
ispartof IEEE signal processing letters, 2017-05, Vol.24 (5), p.609-613
issn 1070-9908
1558-2361
language eng
recordid cdi_ieee_primary_7845566
source IEEE Electronic Library (IEL) Journals
subjects Convolution
Convolutional codes
Convolutional neural network (CNN)
Feature extraction
Image retrieval
multiscale (MS) feature
Neural networks
Robustness
Signal processing algorithms
title MS-RMAC: Multiscale Regional Maximum Activation of Convolutions for Image Retrieval
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-09T08%3A42%3A28IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-crossref_ieee_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=MS-RMAC:%20Multiscale%20Regional%20Maximum%20Activation%20of%20Convolutions%20for%20Image%20Retrieval&rft.jtitle=IEEE%20signal%20processing%20letters&rft.au=Li,%20Yang&rft.date=2017-05&rft.volume=24&rft.issue=5&rft.spage=609&rft.epage=613&rft.pages=609-613&rft.issn=1070-9908&rft.eissn=1558-2361&rft.coden=ISPLEM&rft_id=info:doi/10.1109/LSP.2017.2665522&rft_dat=%3Ccrossref_ieee_%3E10_1109_LSP_2017_2665522%3C/crossref_ieee_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c263t-8663680533f8c4f133ef814943e6795d08ea95bd160688dc5a9b53b66e27e1a3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=7845566&rfr_iscdi=true