GMIF: A Gated Multi-Scale Input Feature Fusion Scheme for Scene Text Detection

The feature fusion of the multi-scale features plays a significant role in localizing text instances of different sizes in the scene text detection (STD) paradigm. The existing approaches are not sufficient to tackle the issues of multi-scale text; consequently, their performance also varies with th...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE access 2022, p.1-1
Hauptverfasser: Ali, Tofik, Siddiqui, Mohammad F. H., Shahab, Sana, Roy, Partha P.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 1
container_issue
container_start_page 1
container_title IEEE access
container_volume
creator Ali, Tofik
Siddiqui, Mohammad F. H.
Shahab, Sana
Roy, Partha P.
description The feature fusion of the multi-scale features plays a significant role in localizing text instances of different sizes in the scene text detection (STD) paradigm. The existing approaches are not sufficient to tackle the issues of multi-scale text; consequently, their performance also varies with the text size. Here we propose a gated multi-scale input feature fusion (GMIF) approach to overcome this issue in STD. The GMIF generates the local features from down-scaled input images and propagates these features from low resolution to the higher resolution global features through a gated recurrent unit-like mechanism. The consistent performance of the GMIF is validated with different text instance sizes of the test-set of the Total-text dataset. The GMIF obtained the performance in range (Precision 88.554-89.106, Recall 85.452-85.790, and f-measures 87.072 - 87.417) with marginal deviation, whereas the current state-of-the-art method, DBNet++, acquired in range (Precision 73.005-82.666, Recall 80.912-87.274, and f-measures 76.755 - 84.183) with significant deviation. Besides this, GMIF also achieved the best performance (f-measures) over ICDAR 2015 (as 88.0), Total-Text (as 87.4), and the second-best over theMSRA-TD500 (as 85.2) dataset.We have conducted an ablation study to show the impact of different components of the GMIF on the STD tasks, which shows the effectiveness of the overall GMIF approach.
doi_str_mv 10.1109/ACCESS.2022.3203691
format Article
fullrecord <record><control><sourceid>ieee</sourceid><recordid>TN_cdi_ieee_primary_9874745</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>9874745</ieee_id><sourcerecordid>9874745</sourcerecordid><originalsourceid>FETCH-ieee_primary_98747453</originalsourceid><addsrcrecordid>eNp9ibkOgkAURScmJhLlC2jeD4CzsNoRBKTABnoywUfEsAWGRP9eCmtPc0_uIcRg1GKMBucwiuKisDjl3BKcCjdgO6Jx5gamcIR7IPqyvOiGv12Op5F7mmfJBUJIpcIH5GunWrOoZYeQDdOqIEGp1hkhWZd2HKCon9gjNOO8KQ4IJb4VXFFhrbZ-IvtGdgvqvz0SI4nL6Ga2iFhNc9vL-VMFvmd7tiP-1y-Bajxi</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>GMIF: A Gated Multi-Scale Input Feature Fusion Scheme for Scene Text Detection</title><source>IEEE Open Access Journals</source><source>DOAJ Directory of Open Access Journals</source><source>EZB-FREE-00999 freely available EZB journals</source><creator>Ali, Tofik ; Siddiqui, Mohammad F. H. ; Shahab, Sana ; Roy, Partha P.</creator><creatorcontrib>Ali, Tofik ; Siddiqui, Mohammad F. H. ; Shahab, Sana ; Roy, Partha P.</creatorcontrib><description>The feature fusion of the multi-scale features plays a significant role in localizing text instances of different sizes in the scene text detection (STD) paradigm. The existing approaches are not sufficient to tackle the issues of multi-scale text; consequently, their performance also varies with the text size. Here we propose a gated multi-scale input feature fusion (GMIF) approach to overcome this issue in STD. The GMIF generates the local features from down-scaled input images and propagates these features from low resolution to the higher resolution global features through a gated recurrent unit-like mechanism. The consistent performance of the GMIF is validated with different text instance sizes of the test-set of the Total-text dataset. The GMIF obtained the performance in range (Precision 88.554-89.106, Recall 85.452-85.790, and f-measures 87.072 - 87.417) with marginal deviation, whereas the current state-of-the-art method, DBNet++, acquired in range (Precision 73.005-82.666, Recall 80.912-87.274, and f-measures 76.755 - 84.183) with significant deviation. Besides this, GMIF also achieved the best performance (f-measures) over ICDAR 2015 (as 88.0), Total-Text (as 87.4), and the second-best over theMSRA-TD500 (as 85.2) dataset.We have conducted an ablation study to show the impact of different components of the GMIF on the STD tasks, which shows the effectiveness of the overall GMIF approach.</description><identifier>EISSN: 2169-3536</identifier><identifier>DOI: 10.1109/ACCESS.2022.3203691</identifier><identifier>CODEN: IAECCG</identifier><language>eng</language><publisher>IEEE</publisher><subject>deep neural networks ; Electronic mail ; Feature extraction ; feature-fusion ; GSM ; Image segmentation ; multi-scale feature ; multi-scale text ; Object detection ; Real-time systems ; Scene text detection ; Testing</subject><ispartof>IEEE access, 2022, p.1-1</ispartof><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><orcidid>0000-0002-5735-5254</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/9874745$$EHTML$$P50$$Gieee$$Hfree_for_read</linktohtml><link.rule.ids>314,777,781,861,4010,27614,27904,27905,27906,54914</link.rule.ids></links><search><creatorcontrib>Ali, Tofik</creatorcontrib><creatorcontrib>Siddiqui, Mohammad F. H.</creatorcontrib><creatorcontrib>Shahab, Sana</creatorcontrib><creatorcontrib>Roy, Partha P.</creatorcontrib><title>GMIF: A Gated Multi-Scale Input Feature Fusion Scheme for Scene Text Detection</title><title>IEEE access</title><addtitle>Access</addtitle><description>The feature fusion of the multi-scale features plays a significant role in localizing text instances of different sizes in the scene text detection (STD) paradigm. The existing approaches are not sufficient to tackle the issues of multi-scale text; consequently, their performance also varies with the text size. Here we propose a gated multi-scale input feature fusion (GMIF) approach to overcome this issue in STD. The GMIF generates the local features from down-scaled input images and propagates these features from low resolution to the higher resolution global features through a gated recurrent unit-like mechanism. The consistent performance of the GMIF is validated with different text instance sizes of the test-set of the Total-text dataset. The GMIF obtained the performance in range (Precision 88.554-89.106, Recall 85.452-85.790, and f-measures 87.072 - 87.417) with marginal deviation, whereas the current state-of-the-art method, DBNet++, acquired in range (Precision 73.005-82.666, Recall 80.912-87.274, and f-measures 76.755 - 84.183) with significant deviation. Besides this, GMIF also achieved the best performance (f-measures) over ICDAR 2015 (as 88.0), Total-Text (as 87.4), and the second-best over theMSRA-TD500 (as 85.2) dataset.We have conducted an ablation study to show the impact of different components of the GMIF on the STD tasks, which shows the effectiveness of the overall GMIF approach.</description><subject>deep neural networks</subject><subject>Electronic mail</subject><subject>Feature extraction</subject><subject>feature-fusion</subject><subject>GSM</subject><subject>Image segmentation</subject><subject>multi-scale feature</subject><subject>multi-scale text</subject><subject>Object detection</subject><subject>Real-time systems</subject><subject>Scene text detection</subject><subject>Testing</subject><issn>2169-3536</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>ESBDL</sourceid><sourceid>RIE</sourceid><recordid>eNp9ibkOgkAURScmJhLlC2jeD4CzsNoRBKTABnoywUfEsAWGRP9eCmtPc0_uIcRg1GKMBucwiuKisDjl3BKcCjdgO6Jx5gamcIR7IPqyvOiGv12Op5F7mmfJBUJIpcIH5GunWrOoZYeQDdOqIEGp1hkhWZd2HKCon9gjNOO8KQ4IJb4VXFFhrbZ-IvtGdgvqvz0SI4nL6Ga2iFhNc9vL-VMFvmd7tiP-1y-Bajxi</recordid><startdate>2022</startdate><enddate>2022</enddate><creator>Ali, Tofik</creator><creator>Siddiqui, Mohammad F. H.</creator><creator>Shahab, Sana</creator><creator>Roy, Partha P.</creator><general>IEEE</general><scope>97E</scope><scope>ESBDL</scope><scope>RIA</scope><scope>RIE</scope><orcidid>https://orcid.org/0000-0002-5735-5254</orcidid></search><sort><creationdate>2022</creationdate><title>GMIF: A Gated Multi-Scale Input Feature Fusion Scheme for Scene Text Detection</title><author>Ali, Tofik ; Siddiqui, Mohammad F. H. ; Shahab, Sana ; Roy, Partha P.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-ieee_primary_98747453</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>deep neural networks</topic><topic>Electronic mail</topic><topic>Feature extraction</topic><topic>feature-fusion</topic><topic>GSM</topic><topic>Image segmentation</topic><topic>multi-scale feature</topic><topic>multi-scale text</topic><topic>Object detection</topic><topic>Real-time systems</topic><topic>Scene text detection</topic><topic>Testing</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Ali, Tofik</creatorcontrib><creatorcontrib>Siddiqui, Mohammad F. H.</creatorcontrib><creatorcontrib>Shahab, Sana</creatorcontrib><creatorcontrib>Roy, Partha P.</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE Open Access Journals</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><jtitle>IEEE access</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Ali, Tofik</au><au>Siddiqui, Mohammad F. H.</au><au>Shahab, Sana</au><au>Roy, Partha P.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>GMIF: A Gated Multi-Scale Input Feature Fusion Scheme for Scene Text Detection</atitle><jtitle>IEEE access</jtitle><stitle>Access</stitle><date>2022</date><risdate>2022</risdate><spage>1</spage><epage>1</epage><pages>1-1</pages><eissn>2169-3536</eissn><coden>IAECCG</coden><abstract>The feature fusion of the multi-scale features plays a significant role in localizing text instances of different sizes in the scene text detection (STD) paradigm. The existing approaches are not sufficient to tackle the issues of multi-scale text; consequently, their performance also varies with the text size. Here we propose a gated multi-scale input feature fusion (GMIF) approach to overcome this issue in STD. The GMIF generates the local features from down-scaled input images and propagates these features from low resolution to the higher resolution global features through a gated recurrent unit-like mechanism. The consistent performance of the GMIF is validated with different text instance sizes of the test-set of the Total-text dataset. The GMIF obtained the performance in range (Precision 88.554-89.106, Recall 85.452-85.790, and f-measures 87.072 - 87.417) with marginal deviation, whereas the current state-of-the-art method, DBNet++, acquired in range (Precision 73.005-82.666, Recall 80.912-87.274, and f-measures 76.755 - 84.183) with significant deviation. Besides this, GMIF also achieved the best performance (f-measures) over ICDAR 2015 (as 88.0), Total-Text (as 87.4), and the second-best over theMSRA-TD500 (as 85.2) dataset.We have conducted an ablation study to show the impact of different components of the GMIF on the STD tasks, which shows the effectiveness of the overall GMIF approach.</abstract><pub>IEEE</pub><doi>10.1109/ACCESS.2022.3203691</doi><orcidid>https://orcid.org/0000-0002-5735-5254</orcidid><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2169-3536
ispartof IEEE access, 2022, p.1-1
issn 2169-3536
language eng
recordid cdi_ieee_primary_9874745
source IEEE Open Access Journals; DOAJ Directory of Open Access Journals; EZB-FREE-00999 freely available EZB journals
subjects deep neural networks
Electronic mail
Feature extraction
feature-fusion
GSM
Image segmentation
multi-scale feature
multi-scale text
Object detection
Real-time systems
Scene text detection
Testing
title GMIF: A Gated Multi-Scale Input Feature Fusion Scheme for Scene Text Detection
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-20T01%3A53%3A49IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=GMIF:%20A%20Gated%20Multi-Scale%20Input%20Feature%20Fusion%20Scheme%20for%20Scene%20Text%20Detection&rft.jtitle=IEEE%20access&rft.au=Ali,%20Tofik&rft.date=2022&rft.spage=1&rft.epage=1&rft.pages=1-1&rft.eissn=2169-3536&rft.coden=IAECCG&rft_id=info:doi/10.1109/ACCESS.2022.3203691&rft_dat=%3Cieee%3E9874745%3C/ieee%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=9874745&rfr_iscdi=true