System for an intelligent office document analysis, recognition and description
Identifieur interne : 001476 ( France/Analysis ); précédent : 001475; suivant : 001477System for an intelligent office document analysis, recognition and description
Auteurs : Philippe Chauvet [France] ; Jaime Lopez-Krahe [France] ; Erik Taflin [France] ; Henri Maître [France]Source :
- Signal Processing [ 0165-1684 ] ; 1993.
Abstract
The authors propose a system for complex-document analysis, coding and archiving. These are achieved using image block segmentation and recognition. This paper describes an advanced document images analysis which involves a multi-layer description of a document and leads to a semantic analysis of its content for an adaptive coding orientation in order to optimize the archiving. It investigates the adaptive aspect that any coding oriented system should now acquire for an intelligent archiving of documents. It is necessary for any intelligent document archiving system to be adaptive to solve the problem of complex-document analysis. Hence, the method presented in this paper consists of document segmentation using a recursive tool based on a run-length smoothing algorithm. This tool performs a pyramidal structure analysis of documents and therefore enables the coding algorithm to adapt to the types of the segmented blocks of document. The segmentation is performed in conjunction with a block recognition system. Recognition is made using a multivariate statistical discriminant analysis with a classification based on linear discriminant functions and on a morphological analysis of the document. It provides an identification of consistent homogeneous blocks of the document: graphics, text blocks, logical inserts, etc. This paper discusses robustness and precision of the segmentation and recognition stages along with experimental results. The classification method yields 97% of correct block classification.
Url:
DOI: 10.1016/0165-1684(93)90041-8
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 001938
- to stream Istex, to step Curation: 001938
- to stream Istex, to step Checkpoint: 001001
- to stream Main, to step Merge: 002263
- to stream Main, to step Curation: 002198
- to stream Main, to step Exploration: 002198
- to stream France, to step Extraction: 001476
Links to Exploration step
ISTEX:C65DA217D5E93B79F5E8B1DA6922ABCEB3277DF2Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title>System for an intelligent office document analysis, recognition and description</title><author><name sortKey="Chauvet, Philippe" sort="Chauvet, Philippe" uniqKey="Chauvet P" first="Philippe" last="Chauvet">Philippe Chauvet</name></author><author><name sortKey="Lopez Krahe, Jaime" sort="Lopez Krahe, Jaime" uniqKey="Lopez Krahe J" first="Jaime" last="Lopez-Krahe">Jaime Lopez-Krahe</name></author><author><name sortKey="Taflin, Erik" sort="Taflin, Erik" uniqKey="Taflin E" first="Erik" last="Taflin">Erik Taflin</name></author><author><name sortKey="Maitre, Henri" sort="Maitre, Henri" uniqKey="Maitre H" first="Henri" last="Maître">Henri Maître</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:C65DA217D5E93B79F5E8B1DA6922ABCEB3277DF2</idno><date when="1993" year="1993">1993</date><idno type="doi">10.1016/0165-1684(93)90041-8</idno><idno type="url">https://api.istex.fr/document/C65DA217D5E93B79F5E8B1DA6922ABCEB3277DF2/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">001938</idno><idno type="wicri:Area/Istex/Curation">001938</idno><idno type="wicri:Area/Istex/Checkpoint">001001</idno><idno type="wicri:doubleKey">0165-1684:1993:Chauvet P:system:for:an</idno><idno type="wicri:Area/Main/Merge">002263</idno><idno type="wicri:Area/Main/Curation">002198</idno><idno type="wicri:Area/Main/Exploration">002198</idno><idno type="wicri:Area/France/Extraction">001476</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a">System for an intelligent office document analysis, recognition and description</title><author><name sortKey="Chauvet, Philippe" sort="Chauvet, Philippe" uniqKey="Chauvet P" first="Philippe" last="Chauvet">Philippe Chauvet</name><affiliation wicri:level="1"><country wicri:rule="url">France</country></affiliation><affiliation wicri:level="3"><country xml:lang="fr">France</country><wicri:regionArea>Département Images, Télécom Paris, 46 rue Barrault, 75634 Paris Cedex 13</wicri:regionArea><placeName><region type="region" nuts="2">Île-de-France</region><settlement type="city">Paris</settlement></placeName></affiliation></author><author><name sortKey="Lopez Krahe, Jaime" sort="Lopez Krahe, Jaime" uniqKey="Lopez Krahe J" first="Jaime" last="Lopez-Krahe">Jaime Lopez-Krahe</name><affiliation wicri:level="3"><country xml:lang="fr">France</country><wicri:regionArea>Département Images, Télécom Paris, 46 rue Barrault, 75634 Paris Cedex 13</wicri:regionArea><placeName><region type="region" nuts="2">Île-de-France</region><settlement type="city">Paris</settlement></placeName></affiliation></author><author><name sortKey="Taflin, Erik" sort="Taflin, Erik" uniqKey="Taflin E" first="Erik" last="Taflin">Erik Taflin</name><affiliation wicri:level="3"><country xml:lang="fr">France</country><wicri:regionArea>UAP, La Mission des Technologies Nouvelles, 20ter rue de Bezons, 92411 Courbevoie Cedex</wicri:regionArea><placeName><region type="region" nuts="2">Île-de-France</region><settlement type="city">Courbevoie</settlement></placeName></affiliation></author><author><name sortKey="Maitre, Henri" sort="Maitre, Henri" uniqKey="Maitre H" first="Henri" last="Maître">Henri Maître</name><affiliation wicri:level="3"><country xml:lang="fr">France</country><wicri:regionArea>Département Images, Télécom Paris, 46 rue Barrault, 75634 Paris Cedex 13</wicri:regionArea><placeName><region type="region" nuts="2">Île-de-France</region><settlement type="city">Paris</settlement></placeName></affiliation></author></analytic><monogr></monogr><series><title level="j">Signal Processing</title><title level="j" type="abbrev">SIGPRO</title><idno type="ISSN">0165-1684</idno><imprint><publisher>ELSEVIER</publisher><date type="published" when="1993">1993</date><biblScope unit="volume">32</biblScope><biblScope unit="issue">1–2</biblScope><biblScope unit="page" from="161">161</biblScope><biblScope unit="page" to="190">190</biblScope></imprint><idno type="ISSN">0165-1684</idno></series><idno type="istex">C65DA217D5E93B79F5E8B1DA6922ABCEB3277DF2</idno><idno type="DOI">10.1016/0165-1684(93)90041-8</idno><idno type="PII">0165-1684(93)90041-8</idno></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">0165-1684</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass><langUsage><language ident="en">en</language></langUsage></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">The authors propose a system for complex-document analysis, coding and archiving. These are achieved using image block segmentation and recognition. This paper describes an advanced document images analysis which involves a multi-layer description of a document and leads to a semantic analysis of its content for an adaptive coding orientation in order to optimize the archiving. It investigates the adaptive aspect that any coding oriented system should now acquire for an intelligent archiving of documents. It is necessary for any intelligent document archiving system to be adaptive to solve the problem of complex-document analysis. Hence, the method presented in this paper consists of document segmentation using a recursive tool based on a run-length smoothing algorithm. This tool performs a pyramidal structure analysis of documents and therefore enables the coding algorithm to adapt to the types of the segmented blocks of document. The segmentation is performed in conjunction with a block recognition system. Recognition is made using a multivariate statistical discriminant analysis with a classification based on linear discriminant functions and on a morphological analysis of the document. It provides an identification of consistent homogeneous blocks of the document: graphics, text blocks, logical inserts, etc. This paper discusses robustness and precision of the segmentation and recognition stages along with experimental results. The classification method yields 97% of correct block classification.</div></front></TEI><affiliations><list><country><li>France</li></country><region><li>Île-de-France</li></region><settlement><li>Courbevoie</li><li>Paris</li></settlement></list><tree><country name="France"><noRegion><name sortKey="Chauvet, Philippe" sort="Chauvet, Philippe" uniqKey="Chauvet P" first="Philippe" last="Chauvet">Philippe Chauvet</name></noRegion><name sortKey="Chauvet, Philippe" sort="Chauvet, Philippe" uniqKey="Chauvet P" first="Philippe" last="Chauvet">Philippe Chauvet</name><name sortKey="Lopez Krahe, Jaime" sort="Lopez Krahe, Jaime" uniqKey="Lopez Krahe J" first="Jaime" last="Lopez-Krahe">Jaime Lopez-Krahe</name><name sortKey="Maitre, Henri" sort="Maitre, Henri" uniqKey="Maitre H" first="Henri" last="Maître">Henri Maître</name><name sortKey="Taflin, Erik" sort="Taflin, Erik" uniqKey="Taflin E" first="Erik" last="Taflin">Erik Taflin</name></country></tree></affiliations></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/France/explor/LeHavreV1/Data/France/Analysis
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 001476 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/France/Analysis/biblio.hfd -nk 001476 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/France
|area= LeHavreV1
|flux= France
|étape= Analysis
|type= RBID
|clé= ISTEX:C65DA217D5E93B79F5E8B1DA6922ABCEB3277DF2
|texte= System for an intelligent office document analysis, recognition and description
}}
| This area was generated with Dilib version V0.6.25. |  |