Vision and language-based multi-modal mixed fusion fine-grained recognition method

The invention provides a vision and language-based multi-modal mixed fusion fine-grained recognition method, and belongs to the technical field of deep learning. The method comprises the following steps: extracting visual features from a visual mode and extracting language features from a language m...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: CHEN YI, ZHU BIN, XIE BO, WANG RUNHUA, ZOU RONGPING, XIA ANNING, YANG HUA
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator CHEN YI
ZHU BIN
XIE BO
WANG RUNHUA
ZOU RONGPING
XIA ANNING
YANG HUA
description The invention provides a vision and language-based multi-modal mixed fusion fine-grained recognition method, and belongs to the technical field of deep learning. The method comprises the following steps: extracting visual features from a visual mode and extracting language features from a language mode by using a feature extraction module; wherein the visual features are fed to a visual modal classifier to determine a visual modal classification result, and the language features are fed to a language modal classifier to obtain a language modal classification result; a feature fusion module is utilized to generate joint features based on the visual features and the language features, the joint features are fed to a multi-head self-attention layer, a feature fusion result is obtained after the joint features pass through a full connection layer, and the classification confidence of the feature fusion result is calculated; and a result fusion module is utilized to determine weights for the classification confide
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN118094172A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN118094172A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN118094172A3</originalsourceid><addsrcrecordid>eNrjZAgKyyzOzM9TSMxLUchJzEsvTUxP1U1KLE5NUcgtzSnJ1M3NT0nMUcjNrACKpJWC1aZl5qXqphclAqkUhaLU5Pz0vMwSkERuaklGfgoPA2taYk5xKi-U5mZQdHMNcfbQTS3Ij08tLkhMTs1LLYl39jM0tDCwNDE0N3I0JkYNAB0VODE</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Vision and language-based multi-modal mixed fusion fine-grained recognition method</title><source>esp@cenet</source><creator>CHEN YI ; ZHU BIN ; XIE BO ; WANG RUNHUA ; ZOU RONGPING ; XIA ANNING ; YANG HUA</creator><creatorcontrib>CHEN YI ; ZHU BIN ; XIE BO ; WANG RUNHUA ; ZOU RONGPING ; XIA ANNING ; YANG HUA</creatorcontrib><description>The invention provides a vision and language-based multi-modal mixed fusion fine-grained recognition method, and belongs to the technical field of deep learning. The method comprises the following steps: extracting visual features from a visual mode and extracting language features from a language mode by using a feature extraction module; wherein the visual features are fed to a visual modal classifier to determine a visual modal classification result, and the language features are fed to a language modal classifier to obtain a language modal classification result; a feature fusion module is utilized to generate joint features based on the visual features and the language features, the joint features are fed to a multi-head self-attention layer, a feature fusion result is obtained after the joint features pass through a full connection layer, and the classification confidence of the feature fusion result is calculated; and a result fusion module is utilized to determine weights for the classification confide</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTING ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PHYSICS</subject><creationdate>2024</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20240528&amp;DB=EPODOC&amp;CC=CN&amp;NR=118094172A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25563,76418</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20240528&amp;DB=EPODOC&amp;CC=CN&amp;NR=118094172A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>CHEN YI</creatorcontrib><creatorcontrib>ZHU BIN</creatorcontrib><creatorcontrib>XIE BO</creatorcontrib><creatorcontrib>WANG RUNHUA</creatorcontrib><creatorcontrib>ZOU RONGPING</creatorcontrib><creatorcontrib>XIA ANNING</creatorcontrib><creatorcontrib>YANG HUA</creatorcontrib><title>Vision and language-based multi-modal mixed fusion fine-grained recognition method</title><description>The invention provides a vision and language-based multi-modal mixed fusion fine-grained recognition method, and belongs to the technical field of deep learning. The method comprises the following steps: extracting visual features from a visual mode and extracting language features from a language mode by using a feature extraction module; wherein the visual features are fed to a visual modal classifier to determine a visual modal classification result, and the language features are fed to a language modal classifier to obtain a language modal classification result; a feature fusion module is utilized to generate joint features based on the visual features and the language features, the joint features are fed to a multi-head self-attention layer, a feature fusion result is obtained after the joint features pass through a full connection layer, and the classification confidence of the feature fusion result is calculated; and a result fusion module is utilized to determine weights for the classification confide</description><subject>CALCULATING</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2024</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZAgKyyzOzM9TSMxLUchJzEsvTUxP1U1KLE5NUcgtzSnJ1M3NT0nMUcjNrACKpJWC1aZl5qXqphclAqkUhaLU5Pz0vMwSkERuaklGfgoPA2taYk5xKi-U5mZQdHMNcfbQTS3Ij08tLkhMTs1LLYl39jM0tDCwNDE0N3I0JkYNAB0VODE</recordid><startdate>20240528</startdate><enddate>20240528</enddate><creator>CHEN YI</creator><creator>ZHU BIN</creator><creator>XIE BO</creator><creator>WANG RUNHUA</creator><creator>ZOU RONGPING</creator><creator>XIA ANNING</creator><creator>YANG HUA</creator><scope>EVB</scope></search><sort><creationdate>20240528</creationdate><title>Vision and language-based multi-modal mixed fusion fine-grained recognition method</title><author>CHEN YI ; ZHU BIN ; XIE BO ; WANG RUNHUA ; ZOU RONGPING ; XIA ANNING ; YANG HUA</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN118094172A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2024</creationdate><topic>CALCULATING</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>CHEN YI</creatorcontrib><creatorcontrib>ZHU BIN</creatorcontrib><creatorcontrib>XIE BO</creatorcontrib><creatorcontrib>WANG RUNHUA</creatorcontrib><creatorcontrib>ZOU RONGPING</creatorcontrib><creatorcontrib>XIA ANNING</creatorcontrib><creatorcontrib>YANG HUA</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>CHEN YI</au><au>ZHU BIN</au><au>XIE BO</au><au>WANG RUNHUA</au><au>ZOU RONGPING</au><au>XIA ANNING</au><au>YANG HUA</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Vision and language-based multi-modal mixed fusion fine-grained recognition method</title><date>2024-05-28</date><risdate>2024</risdate><abstract>The invention provides a vision and language-based multi-modal mixed fusion fine-grained recognition method, and belongs to the technical field of deep learning. The method comprises the following steps: extracting visual features from a visual mode and extracting language features from a language mode by using a feature extraction module; wherein the visual features are fed to a visual modal classifier to determine a visual modal classification result, and the language features are fed to a language modal classifier to obtain a language modal classification result; a feature fusion module is utilized to generate joint features based on the visual features and the language features, the joint features are fed to a multi-head self-attention layer, a feature fusion result is obtained after the joint features pass through a full connection layer, and the classification confidence of the feature fusion result is calculated; and a result fusion module is utilized to determine weights for the classification confide</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN118094172A
source esp@cenet
subjects CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
title Vision and language-based multi-modal mixed fusion fine-grained recognition method
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T15%3A41%3A04IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=CHEN%20YI&rft.date=2024-05-28&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN118094172A%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true