Efficient Language-Independent Retrieval of Printed Documents without OCR
Identifieur interne : 000972 ( Main/Exploration ); précédent : 000971; suivant : 000973Efficient Language-Independent Retrieval of Printed Documents without OCR
Auteurs : Walid Magdy [Irlande (pays)] ; Kareem Darwish [Égypte] ; Motaz El-Saban [Égypte]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2009.
Abstract
Abstract: Recent book digitization initiatives have facilitated the access and search of millions of books. Although OCR remains essential for retrieving printed documents, OCR engines remain limited in the languages they handle and are generally expensive to build. This paper proposes a language independent approach that enables search through printed documents in a way that combines image-based matching with conventional IR techniques without using OCR. While image-based matching can be effective in finding similar words, complementing it with efficient retrieval techniques allows for sub-word matching, term weighting, and document ranking. The basic idea is that similar connected elements in printed documents are clustered and represented with ID’s, which are then used to generate equivalent textual representations. The resultant representations are indexed using an IR engine and searched using the equivalent ID’s of the connected elements in queries. Though, the main benefit of the proposed approach lies in languages for which no OCR exists, the technique was tested on English and Arabic to ascertain the relative effectiveness of the approach. The approach achieves more than 61% relative effectiveness compared to using OCR for both languages. While the reported numbers are lower than that of OCR-based approaches, the proposed method is fully automated, does not require any supervised training, and allows documents to be searchable within a few hours.
Url:
DOI: 10.1007/978-3-642-03784-9_33
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 000001
- to stream Istex, to step Curation: 000001
- to stream Istex, to step Checkpoint: 000494
- to stream Main, to step Merge: 000980
- to stream Main, to step Curation: 000972
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Efficient Language-Independent Retrieval of Printed Documents without OCR</title><author><name sortKey="Magdy, Walid" sort="Magdy, Walid" uniqKey="Magdy W" first="Walid" last="Magdy">Walid Magdy</name></author><author><name sortKey="Darwish, Kareem" sort="Darwish, Kareem" uniqKey="Darwish K" first="Kareem" last="Darwish">Kareem Darwish</name></author><author><name sortKey="El Saban, Motaz" sort="El Saban, Motaz" uniqKey="El Saban M" first="Motaz" last="El-Saban">Motaz El-Saban</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:84B7CE51FF2AA4E37BFDD857BF5CA81233E30FF1</idno><date when="2009" year="2009">2009</date><idno type="doi">10.1007/978-3-642-03784-9_33</idno><idno type="url">https://api.istex.fr/document/84B7CE51FF2AA4E37BFDD857BF5CA81233E30FF1/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000001</idno><idno type="wicri:Area/Istex/Curation">000001</idno><idno type="wicri:Area/Istex/Checkpoint">000494</idno><idno type="wicri:doubleKey">0302-9743:2009:Magdy W:efficient:language:independent</idno><idno type="wicri:Area/Main/Merge">000980</idno><idno type="wicri:Area/Main/Curation">000972</idno><idno type="wicri:Area/Main/Exploration">000972</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Efficient Language-Independent Retrieval of Printed Documents without OCR</title><author><name sortKey="Magdy, Walid" sort="Magdy, Walid" uniqKey="Magdy W" first="Walid" last="Magdy">Walid Magdy</name><affiliation wicri:level="1"><country xml:lang="fr">Irlande (pays)</country><wicri:regionArea>School of Computing, Dublin City University, Dublin 9</wicri:regionArea><wicri:noRegion>Dublin 9</wicri:noRegion></affiliation><affiliation wicri:level="1"><country wicri:rule="url">Irlande (pays)</country></affiliation></author><author><name sortKey="Darwish, Kareem" sort="Darwish, Kareem" uniqKey="Darwish K" first="Kareem" last="Darwish">Kareem Darwish</name><affiliation wicri:level="1"><country xml:lang="fr">Égypte</country><wicri:regionArea>Cairo Microsoft Innovation Center, Microsoft, Smart Village, B115, Abou Rawash</wicri:regionArea><wicri:noRegion>Abou Rawash</wicri:noRegion></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: kareemd@microsoft.com</wicri:noCountry></affiliation></author><author><name sortKey="El Saban, Motaz" sort="El Saban, Motaz" uniqKey="El Saban M" first="Motaz" last="El-Saban">Motaz El-Saban</name><affiliation wicri:level="1"><country xml:lang="fr">Égypte</country><wicri:regionArea>Cairo Microsoft Innovation Center, Microsoft, Smart Village, B115, Abou Rawash</wicri:regionArea><wicri:noRegion>Abou Rawash</wicri:noRegion></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: motazel@microsoft.com</wicri:noCountry></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2009</date></imprint><idno type="ISSN">0302-9743</idno><idno type="eISSN">1611-3349</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">84B7CE51FF2AA4E37BFDD857BF5CA81233E30FF1</idno><idno type="DOI">10.1007/978-3-642-03784-9_33</idno><idno type="ChapterID">33</idno><idno type="ChapterID">Chap33</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: Recent book digitization initiatives have facilitated the access and search of millions of books. Although OCR remains essential for retrieving printed documents, OCR engines remain limited in the languages they handle and are generally expensive to build. This paper proposes a language independent approach that enables search through printed documents in a way that combines image-based matching with conventional IR techniques without using OCR. While image-based matching can be effective in finding similar words, complementing it with efficient retrieval techniques allows for sub-word matching, term weighting, and document ranking. The basic idea is that similar connected elements in printed documents are clustered and represented with ID’s, which are then used to generate equivalent textual representations. The resultant representations are indexed using an IR engine and searched using the equivalent ID’s of the connected elements in queries. Though, the main benefit of the proposed approach lies in languages for which no OCR exists, the technique was tested on English and Arabic to ascertain the relative effectiveness of the approach. The approach achieves more than 61% relative effectiveness compared to using OCR for both languages. While the reported numbers are lower than that of OCR-based approaches, the proposed method is fully automated, does not require any supervised training, and allows documents to be searchable within a few hours.</div></front></TEI><affiliations><list><country><li>Irlande (pays)</li><li>Égypte</li></country></list><tree><country name="Irlande (pays)"><noRegion><name sortKey="Magdy, Walid" sort="Magdy, Walid" uniqKey="Magdy W" first="Walid" last="Magdy">Walid Magdy</name></noRegion><name sortKey="Magdy, Walid" sort="Magdy, Walid" uniqKey="Magdy W" first="Walid" last="Magdy">Walid Magdy</name></country><country name="Égypte"><noRegion><name sortKey="Darwish, Kareem" sort="Darwish, Kareem" uniqKey="Darwish K" first="Kareem" last="Darwish">Kareem Darwish</name></noRegion><name sortKey="El Saban, Motaz" sort="El Saban, Motaz" uniqKey="El Saban M" first="Motaz" last="El-Saban">Motaz El-Saban</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 000972 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 000972 | 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:84B7CE51FF2AA4E37BFDD857BF5CA81233E30FF1
|texte= Efficient Language-Independent Retrieval of Printed Documents without OCR
}}
| This area was generated with Dilib version V0.6.32. |  |