Fully-Pipelining Hardware Implementation of Neural Network for Text-Based Images Retrieval
Identifieur interne : 001071 ( Main/Exploration ); précédent : 001070; suivant : 001072Fully-Pipelining Hardware Implementation of Neural Network for Text-Based Images Retrieval
Auteurs : Dongwuk Kyoung [Corée du Sud] ; Keechul Jung [Corée du Sud]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2006.
Abstract
Abstract: Many hardware implementations cannot execute the software MLPs’ applications using weight of floating-point data, because hardware design of MLPs usually uses fixed-point arithmetic for high speed and small area. The hardware design using fixed-point arithmetic has two important drawbacks which are low accuracy and flexibility. Therefore, we propose a fully-pipelining architecture of MLPs using floating-point arithmetic in order to solve these two problems. Thus our design method can implement the MLPs having the processing speed improved by optimizing the number of hidden nodes in a repeated processing. We apply a software application of MLPs-based text detection that is computed to be 1722120 times for text detection of a 1152×1546 sized image to hardware implementation. Our preliminary result shows a performance enhancement of about eleven times faster using a fully-pipelining architecture than the software application.
Url:
DOI: 10.1007/11760191_196
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 002C84
- to stream Istex, to step Curation: 002A63
- to stream Istex, to step Checkpoint: 000A44
- to stream Main, to step Merge: 001088
- to stream Main, to step Curation: 001071
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Fully-Pipelining Hardware Implementation of Neural Network for Text-Based Images Retrieval</title>
<author><name sortKey="Kyoung, Dongwuk" sort="Kyoung, Dongwuk" uniqKey="Kyoung D" first="Dongwuk" last="Kyoung">Dongwuk Kyoung</name>
</author>
<author><name sortKey="Jung, Keechul" sort="Jung, Keechul" uniqKey="Jung K" first="Keechul" last="Jung">Keechul Jung</name>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">ISTEX</idno>
<idno type="RBID">ISTEX:0A23C4B658AE49C959EA735EA4A4356E0957135E</idno>
<date when="2006" year="2006">2006</date>
<idno type="doi">10.1007/11760191_196</idno>
<idno type="url">https://api.istex.fr/document/0A23C4B658AE49C959EA735EA4A4356E0957135E/fulltext/pdf</idno>
<idno type="wicri:Area/Istex/Corpus">002C84</idno>
<idno type="wicri:Area/Istex/Curation">002A63</idno>
<idno type="wicri:Area/Istex/Checkpoint">000A44</idno>
<idno type="wicri:doubleKey">0302-9743:2006:Kyoung D:fully:pipelining:hardware</idno>
<idno type="wicri:Area/Main/Merge">001088</idno>
<idno type="wicri:Area/Main/Curation">001071</idno>
<idno type="wicri:Area/Main/Exploration">001071</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Fully-Pipelining Hardware Implementation of Neural Network for Text-Based Images Retrieval</title>
<author><name sortKey="Kyoung, Dongwuk" sort="Kyoung, Dongwuk" uniqKey="Kyoung D" first="Dongwuk" last="Kyoung">Dongwuk Kyoung</name>
<affiliation wicri:level="3"><country xml:lang="fr">Corée du Sud</country>
<wicri:regionArea>HCI Lab., College of Information Science, Soongsil University, Seoul</wicri:regionArea>
<placeName><settlement type="city">Séoul</settlement>
</placeName>
</affiliation>
<affiliation wicri:level="1"><country wicri:rule="url">Corée du Sud</country>
</affiliation>
</author>
<author><name sortKey="Jung, Keechul" sort="Jung, Keechul" uniqKey="Jung K" first="Keechul" last="Jung">Keechul Jung</name>
<affiliation wicri:level="3"><country xml:lang="fr">Corée du Sud</country>
<wicri:regionArea>HCI Lab., College of Information Science, Soongsil University, Seoul</wicri:regionArea>
<placeName><settlement type="city">Séoul</settlement>
</placeName>
</affiliation>
<affiliation wicri:level="1"><country wicri:rule="url">Corée du Sud</country>
</affiliation>
</author>
</analytic>
<monogr></monogr>
<series><title level="s">Lecture Notes in Computer Science</title>
<imprint><date>2006</date>
</imprint>
<idno type="ISSN">0302-9743</idno>
<idno type="eISSN">1611-3349</idno>
<idno type="ISSN">0302-9743</idno>
</series>
<idno type="istex">0A23C4B658AE49C959EA735EA4A4356E0957135E</idno>
<idno type="DOI">10.1007/11760191_196</idno>
<idno type="ChapterID">196</idno>
<idno type="ChapterID">Chap196</idno>
</biblStruct>
</sourceDesc>
<seriesStmt><idno type="ISSN">0302-9743</idno>
</seriesStmt>
</fileDesc>
<profileDesc><textClass></textClass>
<langUsage><language ident="en">en</language>
</langUsage>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">Abstract: Many hardware implementations cannot execute the software MLPs’ applications using weight of floating-point data, because hardware design of MLPs usually uses fixed-point arithmetic for high speed and small area. The hardware design using fixed-point arithmetic has two important drawbacks which are low accuracy and flexibility. Therefore, we propose a fully-pipelining architecture of MLPs using floating-point arithmetic in order to solve these two problems. Thus our design method can implement the MLPs having the processing speed improved by optimizing the number of hidden nodes in a repeated processing. We apply a software application of MLPs-based text detection that is computed to be 1722120 times for text detection of a 1152×1546 sized image to hardware implementation. Our preliminary result shows a performance enhancement of about eleven times faster using a fully-pipelining architecture than the software application.</div>
</front>
</TEI>
<affiliations><list><country><li>Corée du Sud</li>
</country>
<settlement><li>Séoul</li>
</settlement>
</list>
<tree><country name="Corée du Sud"><noRegion><name sortKey="Kyoung, Dongwuk" sort="Kyoung, Dongwuk" uniqKey="Kyoung D" first="Dongwuk" last="Kyoung">Dongwuk Kyoung</name>
</noRegion>
<name sortKey="Jung, Keechul" sort="Jung, Keechul" uniqKey="Jung K" first="Keechul" last="Jung">Keechul Jung</name>
<name sortKey="Jung, Keechul" sort="Jung, Keechul" uniqKey="Jung K" first="Keechul" last="Jung">Keechul Jung</name>
<name sortKey="Kyoung, Dongwuk" sort="Kyoung, Dongwuk" uniqKey="Kyoung D" first="Dongwuk" last="Kyoung">Dongwuk Kyoung</name>
</country>
</tree>
</affiliations>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/Main/Exploration
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 001071 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 001071 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Ticri/CIDE |area= OcrV1 |flux= Main |étape= Exploration |type= RBID |clé= ISTEX:0A23C4B658AE49C959EA735EA4A4356E0957135E |texte= Fully-Pipelining Hardware Implementation of Neural Network for Text-Based Images Retrieval }}
This area was generated with Dilib version V0.6.32. |