Optical recognition of handwritten Chinese characters by hierarchical radical matching method
Identifieur interne : 000581 ( Istex/Corpus ); précédent : 000580; suivant : 000582Optical recognition of handwritten Chinese characters by hierarchical radical matching method
Auteurs : An-Bang Wang ; Kuo-Chin FanSource :
- Pattern Recognition [ 0031-3203 ] ; 1999.
Abstract
In this paper, a radical-based OCR system for the recognition of handwritten Chinese characters is proposed. In our approach, a recursive hierarchical scheme is developed to perform radical extraction first. Character features and radical features are then extracted for matching. Last, a hierarchical radical matching scheme is devised to identify the radicals embedded in an input Chinese character and recognize the input character accordingly. Experiments for radical extraction are conducted on 1856 characters. The successful rate of radical extraction is 92.5%. The average time for radical extraction is 0.65 second per character. Experiments for matching process are conducted on two sets: training set and testing set, each set includes 900 characters. The overall recognition rate in our experiments is 98.2 and 80.9% (for training set and testing set, respectively). The average recognition time of our hierarchical radical matching scheme is 0.274s. There are totally 4716 radicals in 1800 characters. In average, one character consists of 2.62 radicals. Each character will match 7.28 radical templates in average. Thus, each radical will match 2.77 radical temples. The experimental results reveal that our proposed method is feasible, flexible, and effective.
Url:
DOI: 10.1016/S0031-3203(99)00207-1
Links to Exploration step
ISTEX:F56C0EB91A1089AD76635B08235F18333AE5E664Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title>Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title><author><name sortKey="Wang, An Bang" sort="Wang, An Bang" uniqKey="Wang A" first="An-Bang" last="Wang">An-Bang Wang</name><affiliation><mods:affiliation>Department of Banking and Insurance, Private Takming Junior College of Commerce, Taiwan, ROC</mods:affiliation></affiliation></author><author><name sortKey="Fan, Kuo Chin" sort="Fan, Kuo Chin" uniqKey="Fan K" first="Kuo-Chin" last="Fan">Kuo-Chin Fan</name><affiliation><mods:affiliation>Institute of Computer Science and Information Engineering, National Central University, Taiwan, ROC</mods:affiliation></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:F56C0EB91A1089AD76635B08235F18333AE5E664</idno><date when="2000" year="2000">2000</date><idno type="doi">10.1016/S0031-3203(99)00207-1</idno><idno type="url">https://api.istex.fr/document/F56C0EB91A1089AD76635B08235F18333AE5E664/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000581</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a">Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title><author><name sortKey="Wang, An Bang" sort="Wang, An Bang" uniqKey="Wang A" first="An-Bang" last="Wang">An-Bang Wang</name><affiliation><mods:affiliation>Department of Banking and Insurance, Private Takming Junior College of Commerce, Taiwan, ROC</mods:affiliation></affiliation></author><author><name sortKey="Fan, Kuo Chin" sort="Fan, Kuo Chin" uniqKey="Fan K" first="Kuo-Chin" last="Fan">Kuo-Chin Fan</name><affiliation><mods:affiliation>Institute of Computer Science and Information Engineering, National Central University, Taiwan, ROC</mods:affiliation></affiliation></author></analytic><monogr></monogr><series><title level="j">Pattern Recognition</title><title level="j" type="abbrev">PR</title><idno type="ISSN">0031-3203</idno><imprint><publisher>ELSEVIER</publisher><date type="published" when="1999">1999</date><biblScope unit="volume">34</biblScope><biblScope unit="issue">1</biblScope><biblScope unit="page" from="15">15</biblScope><biblScope unit="page" to="35">35</biblScope></imprint><idno type="ISSN">0031-3203</idno></series><idno type="istex">F56C0EB91A1089AD76635B08235F18333AE5E664</idno><idno type="DOI">10.1016/S0031-3203(99)00207-1</idno><idno type="PII">S0031-3203(99)00207-1</idno></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">0031-3203</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass><langUsage><language ident="en">en</language></langUsage></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">In this paper, a radical-based OCR system for the recognition of handwritten Chinese characters is proposed. In our approach, a recursive hierarchical scheme is developed to perform radical extraction first. Character features and radical features are then extracted for matching. Last, a hierarchical radical matching scheme is devised to identify the radicals embedded in an input Chinese character and recognize the input character accordingly. Experiments for radical extraction are conducted on 1856 characters. The successful rate of radical extraction is 92.5%. The average time for radical extraction is 0.65 second per character. Experiments for matching process are conducted on two sets: training set and testing set, each set includes 900 characters. The overall recognition rate in our experiments is 98.2 and 80.9% (for training set and testing set, respectively). The average recognition time of our hierarchical radical matching scheme is 0.274s. There are totally 4716 radicals in 1800 characters. In average, one character consists of 2.62 radicals. Each character will match 7.28 radical templates in average. Thus, each radical will match 2.77 radical temples. The experimental results reveal that our proposed method is feasible, flexible, and effective.</div></front></TEI><istex><corpusName>elsevier</corpusName><author><json:item><name>An-Bang Wang</name><affiliations><json:string>Department of Banking and Insurance, Private Takming Junior College of Commerce, Taiwan, ROC</json:string></affiliations></json:item><json:item><name>Kuo-Chin Fan</name><affiliations><json:string>Institute of Computer Science and Information Engineering, National Central University, Taiwan, ROC</json:string></affiliations></json:item></author><subject><json:item><lang><json:string>eng</json:string></lang><value>Radical extraction</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>Character pattern detection</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>Radical matching</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>Knowledge matching</value></json:item></subject><language><json:string>eng</json:string></language><abstract>In this paper, a radical-based OCR system for the recognition of handwritten Chinese characters is proposed. In our approach, a recursive hierarchical scheme is developed to perform radical extraction first. Character features and radical features are then extracted for matching. Last, a hierarchical radical matching scheme is devised to identify the radicals embedded in an input Chinese character and recognize the input character accordingly. Experiments for radical extraction are conducted on 1856 characters. The successful rate of radical extraction is 92.5%. The average time for radical extraction is 0.65 second per character. Experiments for matching process are conducted on two sets: training set and testing set, each set includes 900 characters. The overall recognition rate in our experiments is 98.2 and 80.9% (for training set and testing set, respectively). The average recognition time of our hierarchical radical matching scheme is 0.274s. There are totally 4716 radicals in 1800 characters. In average, one character consists of 2.62 radicals. Each character will match 7.28 radical templates in average. Thus, each radical will match 2.77 radical temples. The experimental results reveal that our proposed method is feasible, flexible, and effective.</abstract><qualityIndicators><score>7.244</score><pdfVersion>1.2</pdfVersion><pdfPageSize>504 x 714 pts</pdfPageSize><refBibsNative>true</refBibsNative><keywordCount>4</keywordCount><abstractCharCount>1275</abstractCharCount><pdfWordCount>8752</pdfWordCount><pdfCharCount>49936</pdfCharCount><pdfPageCount>21</pdfPageCount><abstractWordCount>187</abstractWordCount></qualityIndicators><title>Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title><pii><json:string>S0031-3203(99)00207-1</json:string></pii><genre><json:string>research-article</json:string></genre><host><volume>34</volume><pii><json:string>S0031-3203(00)X0072-6</json:string></pii><pages><last>35</last><first>15</first></pages><issn><json:string>0031-3203</json:string></issn><issue>1</issue><genre><json:string>Journal</json:string></genre><language><json:string>unknown</json:string></language><title>Pattern Recognition</title><publicationDate>2001</publicationDate></host><categories><wos><json:string>COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE</json:string><json:string>ENGINEERING, ELECTRICAL & ELECTRONIC</json:string></wos></categories><publicationDate>1999</publicationDate><copyrightDate>2000</copyrightDate><doi><json:string>10.1016/S0031-3203(99)00207-1</json:string></doi><id>F56C0EB91A1089AD76635B08235F18333AE5E664</id><fulltext><json:item><original>true</original><mimetype>application/pdf</mimetype><extension>pdf</extension><uri>https://api.istex.fr/document/F56C0EB91A1089AD76635B08235F18333AE5E664/fulltext/pdf</uri></json:item><json:item><original>true</original><mimetype>text/plain</mimetype><extension>txt</extension><uri>https://api.istex.fr/document/F56C0EB91A1089AD76635B08235F18333AE5E664/fulltext/txt</uri></json:item><json:item><original>false</original><mimetype>application/zip</mimetype><extension>zip</extension><uri>https://api.istex.fr/document/F56C0EB91A1089AD76635B08235F18333AE5E664/fulltext/zip</uri></json:item><istex:fulltextTEI uri="https://api.istex.fr/document/F56C0EB91A1089AD76635B08235F18333AE5E664/fulltext/tei"><teiHeader><fileDesc><titleStmt><title level="a">Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title></titleStmt><publicationStmt><authority>ISTEX</authority><publisher>ELSEVIER</publisher><availability><p>ELSEVIER</p></availability><date>2000</date></publicationStmt><notesStmt><note>This work was supported by National Science Council of Taiwan under grant NSC 87-2213-E-008-009.</note><note type="content">Fig. 1: System diagram of our proposed radical-based OCR system.</note><note type="content">Fig. 2: Block diagram of the proposed recursive hierarchical radical extraction module.</note><note type="content">Fig. 3: Ten patterns of Chinese characters.</note><note type="content">Fig. 4: Examples illustrating each pattern of Chinese characters.</note><note type="content">Fig. 5: Detection of component “ ” for characters with SU type.</note><note type="content">Fig. 6: Characters with LUR type are classified into four classes: (a) detection of component “ ”, (b) detection of components “ ” and “ ”, (c) detection of components “ ” and “ ”, (d) detection of component “ ”.</note><note type="content">Fig. 7: Example illustrating how the exception rules are used to exclude the incorrect radical component.</note><note type="content">Fig. 8: Example illustrating radical extraction by making use of straight cut line. Lines with two arrows (such as L1 and L3) are clear straight cut lines. Line with one arrow (such as L2) is nearly straight cut line. The character “ ” are decomposed into four areas by these three vertical cut lines.</note><note type="content">Fig. 9: Example illustrating radical extraction when both horizontal and vertical gaps exist.</note><note type="content">Fig. 10: Block diagram of stroke clustering layer.</note><note type="content">Fig. 11: Illustration of stroke clustering for characters with LR and UD patterns.</note><note type="content">Fig. 12: Redistributing results for left-right characters.</note><note type="content">Fig. 13: Illustration of distance computation for character with LR pattern.</note><note type="content">Fig. 14: The 19 appearing types of a radical.</note><note type="content">Fig. 15: (a) The input character, (b) the process of radical extraction from top to bottom where leaf nodes represent the extracted radicals, (c) the link list formed by the extracted radicals.</note><note type="content">Fig. 16: Examples illustrating radical extraction result: (a)–(f) are characters with UL, UR, LD, LUR, LUR, and SU types, respectively.</note><note type="content">Fig. 17: The block diagram of proposed hierarchical radical matching scheme.</note><note type="content">Fig. 18: (a) The input character “ ”. (b) The three extracted radicals after radical extraction. (c) The templates with which each extracted radical will match. (d) The candidate set after radical matching.</note><note type="content">Fig. 19: The structure of knowledge database.</note><note type="content">Fig. 20: Illustration of the whole character matching “ ” with “ ”.</note><note type="content">Fig. 21: The first sample set contains 300 characters which were written (under constraint) by six writers to form 1800 samples.</note><note type="content">Fig. 22: The 32 radicals used in creating radical database.</note><note type="content">Fig. 23: The second sample set contains 156 characters.</note><note type="content">Table 1:</note><note type="content">Table 2:</note><note type="content">Table 3:</note><note type="content">Table 4: The tabulated results of radical extraction</note><note type="content">Table 5: The tabulated results of hierarchical radical matching scheme (for training set)</note><note type="content">Table 6: The tabulated results of hierarchical radical matching scheme (for testing set)</note><note type="content">Table 7: The recognition speed of hierarchical radical matching scheme</note></notesStmt><sourceDesc><biblStruct type="inbook"><analytic><title level="a">Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title><author><persName><forename type="first">An-Bang</forename><surname>Wang</surname></persName><affiliation>Department of Banking and Insurance, Private Takming Junior College of Commerce, Taiwan, ROC</affiliation></author><author><persName><forename type="first">Kuo-Chin</forename><surname>Fan</surname></persName><note type="correspondence"><p>Corresponding author. Tel.: (03) 4227151-4453; fax: (03) 4222681</p></note><affiliation>Institute of Computer Science and Information Engineering, National Central University, Taiwan, ROC</affiliation></author></analytic><monogr><title level="j">Pattern Recognition</title><title level="j" type="abbrev">PR</title><idno type="pISSN">0031-3203</idno><idno type="PII">S0031-3203(00)X0072-6</idno><imprint><publisher>ELSEVIER</publisher><date type="published" when="1999"></date><biblScope unit="volume">34</biblScope><biblScope unit="issue">1</biblScope><biblScope unit="page" from="15">15</biblScope><biblScope unit="page" to="35">35</biblScope></imprint></monogr><idno type="istex">F56C0EB91A1089AD76635B08235F18333AE5E664</idno><idno type="DOI">10.1016/S0031-3203(99)00207-1</idno><idno type="PII">S0031-3203(99)00207-1</idno></biblStruct></sourceDesc></fileDesc><profileDesc><creation><date>2000</date></creation><langUsage><language ident="en">en</language></langUsage><abstract xml:lang="en"><p>In this paper, a radical-based OCR system for the recognition of handwritten Chinese characters is proposed. In our approach, a recursive hierarchical scheme is developed to perform radical extraction first. Character features and radical features are then extracted for matching. Last, a hierarchical radical matching scheme is devised to identify the radicals embedded in an input Chinese character and recognize the input character accordingly. Experiments for radical extraction are conducted on 1856 characters. The successful rate of radical extraction is 92.5%. The average time for radical extraction is 0.65 second per character. Experiments for matching process are conducted on two sets: training set and testing set, each set includes 900 characters. The overall recognition rate in our experiments is 98.2 and 80.9% (for training set and testing set, respectively). The average recognition time of our hierarchical radical matching scheme is 0.274s. There are totally 4716 radicals in 1800 characters. In average, one character consists of 2.62 radicals. Each character will match 7.28 radical templates in average. Thus, each radical will match 2.77 radical temples. The experimental results reveal that our proposed method is feasible, flexible, and effective.</p></abstract><textClass><keywords scheme="keyword"><list><head>Keywords</head><item><term>Radical extraction</term></item><item><term>Character pattern detection</term></item><item><term>Radical matching</term></item><item><term>Knowledge matching</term></item></list></keywords></textClass></profileDesc><revisionDesc><change when="1999-10-08">Registration</change><change when="1999">Published</change></revisionDesc></teiHeader></istex:fulltextTEI></fulltext><metadata><istex:metadataXml wicri:clean="Elsevier, elements deleted: ce:floats; body; tail"><istex:xmlDeclaration>version="1.0" encoding="utf-8"</istex:xmlDeclaration><istex:docType PUBLIC="-//ES//DTD journal article DTD version 4.5.2//EN//XML" URI="art452.dtd" name="istex:docType"><istex:entity SYSTEM="gr1" NDATA="IMAGE" name="gr1"></istex:entity><istex:entity SYSTEM="gr2" NDATA="IMAGE" name="gr2"></istex:entity><istex:entity SYSTEM="gr3" NDATA="IMAGE" name="gr3"></istex:entity><istex:entity SYSTEM="gr4" NDATA="IMAGE" name="gr4"></istex:entity><istex:entity SYSTEM="fx1" NDATA="IMAGE" name="fx1"></istex:entity><istex:entity SYSTEM="fx2" NDATA="IMAGE" name="fx2"></istex:entity><istex:entity SYSTEM="fx3" NDATA="IMAGE" name="fx3"></istex:entity><istex:entity SYSTEM="fx12" NDATA="IMAGE" name="fx12"></istex:entity><istex:entity SYSTEM="fx5" NDATA="IMAGE" name="fx5"></istex:entity><istex:entity SYSTEM="fx6" NDATA="IMAGE" name="fx6"></istex:entity><istex:entity SYSTEM="fx7" NDATA="IMAGE" name="fx7"></istex:entity><istex:entity SYSTEM="fx8" NDATA="IMAGE" name="fx8"></istex:entity><istex:entity SYSTEM="fx9" NDATA="IMAGE" name="fx9"></istex:entity><istex:entity SYSTEM="fx10" NDATA="IMAGE" name="fx10"></istex:entity><istex:entity SYSTEM="fx11" NDATA="IMAGE" name="fx11"></istex:entity><istex:entity SYSTEM="fx13" NDATA="IMAGE" name="fx13"></istex:entity><istex:entity SYSTEM="fx14" NDATA="IMAGE" name="fx14"></istex:entity><istex:entity SYSTEM="fx15" NDATA="IMAGE" name="fx15"></istex:entity><istex:entity SYSTEM="fx16" NDATA="IMAGE" name="fx16"></istex:entity><istex:entity SYSTEM="fx17" NDATA="IMAGE" name="fx17"></istex:entity><istex:entity SYSTEM="fx18" NDATA="IMAGE" name="fx18"></istex:entity><istex:entity SYSTEM="fx19" NDATA="IMAGE" name="fx19"></istex:entity><istex:entity SYSTEM="fx20" NDATA="IMAGE" name="fx20"></istex:entity><istex:entity SYSTEM="fx21" NDATA="IMAGE" name="fx21"></istex:entity><istex:entity SYSTEM="fx22" NDATA="IMAGE" name="fx22"></istex:entity><istex:entity SYSTEM="fx23" NDATA="IMAGE" name="fx23"></istex:entity><istex:entity SYSTEM="fx24" NDATA="IMAGE" name="fx24"></istex:entity><istex:entity SYSTEM="fx25" NDATA="IMAGE" name="fx25"></istex:entity><istex:entity SYSTEM="fx26" NDATA="IMAGE" name="fx26"></istex:entity><istex:entity SYSTEM="fx27" NDATA="IMAGE" name="fx27"></istex:entity><istex:entity SYSTEM="fx28" NDATA="IMAGE" name="fx28"></istex:entity><istex:entity SYSTEM="fx29" NDATA="IMAGE" name="fx29"></istex:entity><istex:entity SYSTEM="fx30" NDATA="IMAGE" name="fx30"></istex:entity><istex:entity SYSTEM="fx31" NDATA="IMAGE" name="fx31"></istex:entity><istex:entity SYSTEM="fx32" NDATA="IMAGE" name="fx32"></istex:entity><istex:entity SYSTEM="fx33" NDATA="IMAGE" name="fx33"></istex:entity><istex:entity SYSTEM="fx34" NDATA="IMAGE" name="fx34"></istex:entity><istex:entity SYSTEM="fx35" NDATA="IMAGE" name="fx35"></istex:entity><istex:entity SYSTEM="fx36" NDATA="IMAGE" name="fx36"></istex:entity><istex:entity SYSTEM="fx37" NDATA="IMAGE" name="fx37"></istex:entity><istex:entity SYSTEM="fx38" NDATA="IMAGE" name="fx38"></istex:entity><istex:entity SYSTEM="fx39" NDATA="IMAGE" name="fx39"></istex:entity><istex:entity SYSTEM="fx40" NDATA="IMAGE" name="fx40"></istex:entity><istex:entity SYSTEM="fx41" NDATA="IMAGE" name="fx41"></istex:entity><istex:entity SYSTEM="fx42" NDATA="IMAGE" name="fx42"></istex:entity><istex:entity SYSTEM="fx43" NDATA="IMAGE" name="fx43"></istex:entity><istex:entity SYSTEM="fx44" NDATA="IMAGE" name="fx44"></istex:entity><istex:entity SYSTEM="fx45" NDATA="IMAGE" name="fx45"></istex:entity><istex:entity SYSTEM="fx46" NDATA="IMAGE" name="fx46"></istex:entity><istex:entity SYSTEM="fx47" NDATA="IMAGE" name="fx47"></istex:entity><istex:entity SYSTEM="fx48" NDATA="IMAGE" name="fx48"></istex:entity><istex:entity SYSTEM="fx49" NDATA="IMAGE" name="fx49"></istex:entity><istex:entity SYSTEM="fx50" NDATA="IMAGE" name="fx50"></istex:entity><istex:entity SYSTEM="fx51" NDATA="IMAGE" name="fx51"></istex:entity><istex:entity SYSTEM="fx52" NDATA="IMAGE" name="fx52"></istex:entity><istex:entity SYSTEM="fx53" NDATA="IMAGE" name="fx53"></istex:entity><istex:entity SYSTEM="fx54" NDATA="IMAGE" name="fx54"></istex:entity><istex:entity SYSTEM="fx55" NDATA="IMAGE" name="fx55"></istex:entity><istex:entity SYSTEM="fx56" NDATA="IMAGE" name="fx56"></istex:entity><istex:entity SYSTEM="fx57" NDATA="IMAGE" name="fx57"></istex:entity><istex:entity SYSTEM="fx58" NDATA="IMAGE" name="fx58"></istex:entity><istex:entity SYSTEM="fx59" NDATA="IMAGE" name="fx59"></istex:entity><istex:entity SYSTEM="fx60" NDATA="IMAGE" name="fx60"></istex:entity><istex:entity SYSTEM="fx61" NDATA="IMAGE" name="fx61"></istex:entity><istex:entity SYSTEM="fx62" NDATA="IMAGE" name="fx62"></istex:entity><istex:entity SYSTEM="fx63" NDATA="IMAGE" name="fx63"></istex:entity><istex:entity SYSTEM="fx64" NDATA="IMAGE" name="fx64"></istex:entity><istex:entity SYSTEM="fx65" NDATA="IMAGE" name="fx65"></istex:entity><istex:entity SYSTEM="fx66" NDATA="IMAGE" name="fx66"></istex:entity><istex:entity SYSTEM="fx67" NDATA="IMAGE" name="fx67"></istex:entity><istex:entity SYSTEM="fx68" NDATA="IMAGE" name="fx68"></istex:entity><istex:entity SYSTEM="fx69" NDATA="IMAGE" name="fx69"></istex:entity><istex:entity SYSTEM="fx70" NDATA="IMAGE" name="fx70"></istex:entity><istex:entity SYSTEM="fx71" NDATA="IMAGE" name="fx71"></istex:entity><istex:entity SYSTEM="fx72" NDATA="IMAGE" name="fx72"></istex:entity><istex:entity SYSTEM="fx73" NDATA="IMAGE" name="fx73"></istex:entity><istex:entity SYSTEM="fx74" NDATA="IMAGE" name="fx74"></istex:entity><istex:entity SYSTEM="fx75" NDATA="IMAGE" name="fx75"></istex:entity><istex:entity SYSTEM="fx76" NDATA="IMAGE" name="fx76"></istex:entity><istex:entity SYSTEM="fx77" NDATA="IMAGE" name="fx77"></istex:entity><istex:entity SYSTEM="fx78" NDATA="IMAGE" name="fx78"></istex:entity><istex:entity SYSTEM="fx79" NDATA="IMAGE" name="fx79"></istex:entity><istex:entity SYSTEM="fx80" NDATA="IMAGE" name="fx80"></istex:entity><istex:entity SYSTEM="fx81" NDATA="IMAGE" name="fx81"></istex:entity><istex:entity SYSTEM="fx82" NDATA="IMAGE" name="fx82"></istex:entity><istex:entity SYSTEM="fx83" NDATA="IMAGE" name="fx83"></istex:entity><istex:entity SYSTEM="fx84" NDATA="IMAGE" name="fx84"></istex:entity><istex:entity SYSTEM="fx85" NDATA="IMAGE" name="fx85"></istex:entity><istex:entity SYSTEM="fx86" NDATA="IMAGE" name="fx86"></istex:entity><istex:entity SYSTEM="fx87" NDATA="IMAGE" name="fx87"></istex:entity><istex:entity SYSTEM="fx88" NDATA="IMAGE" name="fx88"></istex:entity><istex:entity SYSTEM="fx89" NDATA="IMAGE" name="fx89"></istex:entity><istex:entity SYSTEM="fx90" NDATA="IMAGE" name="fx90"></istex:entity><istex:entity SYSTEM="fx4" NDATA="IMAGE" name="fx4"></istex:entity><istex:entity SYSTEM="gr5" NDATA="IMAGE" name="gr5"></istex:entity><istex:entity SYSTEM="gr6" NDATA="IMAGE" name="gr6"></istex:entity><istex:entity SYSTEM="fx91" NDATA="IMAGE" name="fx91"></istex:entity><istex:entity SYSTEM="fx92" NDATA="IMAGE" name="fx92"></istex:entity><istex:entity SYSTEM="gr7" NDATA="IMAGE" name="gr7"></istex:entity><istex:entity SYSTEM="fx93" NDATA="IMAGE" name="fx93"></istex:entity><istex:entity SYSTEM="gr8" NDATA="IMAGE" name="gr8"></istex:entity><istex:entity SYSTEM="fx94" NDATA="IMAGE" name="fx94"></istex:entity><istex:entity SYSTEM="fx95" NDATA="IMAGE" name="fx95"></istex:entity><istex:entity SYSTEM="fx96" NDATA="IMAGE" name="fx96"></istex:entity><istex:entity SYSTEM="fx97" NDATA="IMAGE" name="fx97"></istex:entity><istex:entity SYSTEM="gr9" NDATA="IMAGE" name="gr9"></istex:entity><istex:entity SYSTEM="fx98" NDATA="IMAGE" name="fx98"></istex:entity><istex:entity SYSTEM="fx99" NDATA="IMAGE" name="fx99"></istex:entity><istex:entity SYSTEM="fx100" NDATA="IMAGE" name="fx100"></istex:entity><istex:entity SYSTEM="fx101" NDATA="IMAGE" name="fx101"></istex:entity><istex:entity SYSTEM="fx102" NDATA="IMAGE" name="fx102"></istex:entity><istex:entity SYSTEM="gr10" NDATA="IMAGE" name="gr10"></istex:entity><istex:entity SYSTEM="gr11" NDATA="IMAGE" name="gr11"></istex:entity><istex:entity SYSTEM="fx103" NDATA="IMAGE" name="fx103"></istex:entity><istex:entity SYSTEM="fx104" NDATA="IMAGE" name="fx104"></istex:entity><istex:entity SYSTEM="gr12" NDATA="IMAGE" name="gr12"></istex:entity><istex:entity SYSTEM="gr13" NDATA="IMAGE" name="gr13"></istex:entity><istex:entity SYSTEM="gr14" NDATA="IMAGE" name="gr14"></istex:entity><istex:entity SYSTEM="fx105" NDATA="IMAGE" name="fx105"></istex:entity><istex:entity SYSTEM="fx106" NDATA="IMAGE" name="fx106"></istex:entity><istex:entity SYSTEM="fx107" NDATA="IMAGE" name="fx107"></istex:entity><istex:entity SYSTEM="gr15" NDATA="IMAGE" name="gr15"></istex:entity><istex:entity SYSTEM="fx108" NDATA="IMAGE" name="fx108"></istex:entity><istex:entity SYSTEM="gr16" NDATA="IMAGE" name="gr16"></istex:entity><istex:entity SYSTEM="gr17" NDATA="IMAGE" name="gr17"></istex:entity><istex:entity SYSTEM="fx109" NDATA="IMAGE" name="fx109"></istex:entity><istex:entity SYSTEM="fx110" NDATA="IMAGE" name="fx110"></istex:entity><istex:entity SYSTEM="fx111" NDATA="IMAGE" name="fx111"></istex:entity><istex:entity SYSTEM="gr18" NDATA="IMAGE" name="gr18"></istex:entity><istex:entity SYSTEM="fx112" NDATA="IMAGE" name="fx112"></istex:entity><istex:entity SYSTEM="fx113" NDATA="IMAGE" name="fx113"></istex:entity><istex:entity SYSTEM="fx114" NDATA="IMAGE" name="fx114"></istex:entity><istex:entity SYSTEM="fx115" NDATA="IMAGE" name="fx115"></istex:entity><istex:entity SYSTEM="fx116" NDATA="IMAGE" name="fx116"></istex:entity><istex:entity SYSTEM="fx117" NDATA="IMAGE" name="fx117"></istex:entity><istex:entity SYSTEM="fx118" NDATA="IMAGE" name="fx118"></istex:entity><istex:entity SYSTEM="fx119" NDATA="IMAGE" name="fx119"></istex:entity><istex:entity SYSTEM="gr19" NDATA="IMAGE" name="gr19"></istex:entity><istex:entity SYSTEM="fx120" NDATA="IMAGE" name="fx120"></istex:entity><istex:entity SYSTEM="fx121" NDATA="IMAGE" name="fx121"></istex:entity><istex:entity SYSTEM="gr20" NDATA="IMAGE" name="gr20"></istex:entity><istex:entity SYSTEM="gr21" NDATA="IMAGE" name="gr21"></istex:entity><istex:entity SYSTEM="gr22" NDATA="IMAGE" name="gr22"></istex:entity><istex:entity SYSTEM="gr23" NDATA="IMAGE" name="gr23"></istex:entity></istex:docType><istex:document><converted-article version="4.5.2" docsubtype="fla"><item-info><jid>PR</jid><aid>1208</aid><ce:pii>S0031-3203(99)00207-1</ce:pii><ce:doi>10.1016/S0031-3203(99)00207-1</ce:doi><ce:copyright type="society" year="2000">Pattern Recognition Society</ce:copyright></item-info><head><ce:article-footnote><ce:label>☆</ce:label><ce:note-para>This work was supported by National Science Council of Taiwan under grant NSC 87-2213-E-008-009.</ce:note-para></ce:article-footnote><ce:title>Optical recognition of handwritten Chinese characters by hierarchical radical matching method</ce:title><ce:author-group><ce:author><ce:given-name>An-Bang</ce:given-name><ce:surname>Wang</ce:surname><ce:cross-ref refid="AUT2">*</ce:cross-ref><ce:cross-ref refid="ORFA"><ce:sup>a</ce:sup></ce:cross-ref></ce:author><ce:author><ce:given-name>Kuo-Chin</ce:given-name><ce:surname>Fan</ce:surname><ce:cross-ref refid="AUT3">*</ce:cross-ref><ce:cross-ref refid="ORFB"><ce:sup>b</ce:sup></ce:cross-ref><ce:cross-ref refid="CORR1">*</ce:cross-ref></ce:author><ce:affiliation id="ORFA"><ce:label>a</ce:label><ce:textfn>Department of Banking and Insurance, Private Takming Junior College of Commerce, Taiwan, ROC</ce:textfn></ce:affiliation><ce:affiliation id="ORFB"><ce:label>b</ce:label><ce:textfn>Institute of Computer Science and Information Engineering, National Central University, Taiwan, ROC</ce:textfn></ce:affiliation><ce:correspondence id="CORR1"><ce:label>*</ce:label><ce:text>Corresponding author. Tel.: (03) 4227151-4453; fax: (03) 4222681</ce:text></ce:correspondence></ce:author-group><ce:date-received day="8" month="10" year="1997"></ce:date-received><ce:date-accepted day="8" month="10" year="1999"></ce:date-accepted><ce:abstract><ce:section-title>Abstract</ce:section-title><ce:abstract-sec><ce:simple-para>In this paper, a radical-based OCR system for the recognition of handwritten Chinese characters is proposed. In our approach, a recursive hierarchical scheme is developed to perform radical extraction first. Character features and radical features are then extracted for matching. Last, a hierarchical radical matching scheme is devised to identify the radicals embedded in an input Chinese character and recognize the input character accordingly. Experiments for radical extraction are conducted on 1856 characters. The successful rate of radical extraction is 92.5%. The average time for radical extraction is 0.65 second per character. Experiments for matching process are conducted on two sets: training set and testing set, each set includes 900 characters. The overall recognition rate in our experiments is 98.2 and 80.9% (for training set and testing set, respectively). The average recognition time of our hierarchical radical matching scheme is 0.274<ce:hsp sp="0.25"></ce:hsp>s. There are totally 4716 radicals in 1800 characters. In average, one character consists of 2.62 radicals. Each character will match 7.28 radical templates in average. Thus, each radical will match 2.77 radical temples. The experimental results reveal that our proposed method is feasible, flexible, and effective.</ce:simple-para></ce:abstract-sec></ce:abstract><ce:keywords class="keyword"><ce:section-title>Keywords</ce:section-title><ce:keyword><ce:text>Radical extraction</ce:text></ce:keyword><ce:keyword><ce:text>Character pattern detection</ce:text></ce:keyword><ce:keyword><ce:text>Radical matching</ce:text></ce:keyword><ce:keyword><ce:text>Knowledge matching</ce:text></ce:keyword></ce:keywords></head></converted-article></istex:document></istex:metadataXml><mods version="3.6"><titleInfo><title>Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title></titleInfo><titleInfo type="alternative" contentType="CDATA"><title>Optical recognition of handwritten Chinese characters by hierarchical radical matching method</title></titleInfo><name type="personal"><namePart type="given">An-Bang</namePart><namePart type="family">Wang</namePart><affiliation>Department of Banking and Insurance, Private Takming Junior College of Commerce, Taiwan, ROC</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Kuo-Chin</namePart><namePart type="family">Fan</namePart><affiliation>Institute of Computer Science and Information Engineering, National Central University, Taiwan, ROC</affiliation><description>Corresponding author. Tel.: (03) 4227151-4453; fax: (03) 4222681</description><role><roleTerm type="text">author</roleTerm></role></name><typeOfResource>text</typeOfResource><genre type="research-article" displayLabel="Full-length article"></genre><originInfo><publisher>ELSEVIER</publisher><dateIssued encoding="w3cdtf">1999</dateIssued><dateValid encoding="w3cdtf">1999-10-08</dateValid><copyrightDate encoding="w3cdtf">2000</copyrightDate></originInfo><language><languageTerm type="code" authority="iso639-2b">eng</languageTerm><languageTerm type="code" authority="rfc3066">en</languageTerm></language><physicalDescription><internetMediaType>text/html</internetMediaType></physicalDescription><abstract lang="en">In this paper, a radical-based OCR system for the recognition of handwritten Chinese characters is proposed. In our approach, a recursive hierarchical scheme is developed to perform radical extraction first. Character features and radical features are then extracted for matching. Last, a hierarchical radical matching scheme is devised to identify the radicals embedded in an input Chinese character and recognize the input character accordingly. Experiments for radical extraction are conducted on 1856 characters. The successful rate of radical extraction is 92.5%. The average time for radical extraction is 0.65 second per character. Experiments for matching process are conducted on two sets: training set and testing set, each set includes 900 characters. The overall recognition rate in our experiments is 98.2 and 80.9% (for training set and testing set, respectively). The average recognition time of our hierarchical radical matching scheme is 0.274s. There are totally 4716 radicals in 1800 characters. In average, one character consists of 2.62 radicals. Each character will match 7.28 radical templates in average. Thus, each radical will match 2.77 radical temples. The experimental results reveal that our proposed method is feasible, flexible, and effective.</abstract><note>This work was supported by National Science Council of Taiwan under grant NSC 87-2213-E-008-009.</note><note type="content">Fig. 1: System diagram of our proposed radical-based OCR system.</note><note type="content">Fig. 2: Block diagram of the proposed recursive hierarchical radical extraction module.</note><note type="content">Fig. 3: Ten patterns of Chinese characters.</note><note type="content">Fig. 4: Examples illustrating each pattern of Chinese characters.</note><note type="content">Fig. 5: Detection of component “ ” for characters with SU type.</note><note type="content">Fig. 6: Characters with LUR type are classified into four classes: (a) detection of component “ ”, (b) detection of components “ ” and “ ”, (c) detection of components “ ” and “ ”, (d) detection of component “ ”.</note><note type="content">Fig. 7: Example illustrating how the exception rules are used to exclude the incorrect radical component.</note><note type="content">Fig. 8: Example illustrating radical extraction by making use of straight cut line. Lines with two arrows (such as L1 and L3) are clear straight cut lines. Line with one arrow (such as L2) is nearly straight cut line. The character “ ” are decomposed into four areas by these three vertical cut lines.</note><note type="content">Fig. 9: Example illustrating radical extraction when both horizontal and vertical gaps exist.</note><note type="content">Fig. 10: Block diagram of stroke clustering layer.</note><note type="content">Fig. 11: Illustration of stroke clustering for characters with LR and UD patterns.</note><note type="content">Fig. 12: Redistributing results for left-right characters.</note><note type="content">Fig. 13: Illustration of distance computation for character with LR pattern.</note><note type="content">Fig. 14: The 19 appearing types of a radical.</note><note type="content">Fig. 15: (a) The input character, (b) the process of radical extraction from top to bottom where leaf nodes represent the extracted radicals, (c) the link list formed by the extracted radicals.</note><note type="content">Fig. 16: Examples illustrating radical extraction result: (a)–(f) are characters with UL, UR, LD, LUR, LUR, and SU types, respectively.</note><note type="content">Fig. 17: The block diagram of proposed hierarchical radical matching scheme.</note><note type="content">Fig. 18: (a) The input character “ ”. (b) The three extracted radicals after radical extraction. (c) The templates with which each extracted radical will match. (d) The candidate set after radical matching.</note><note type="content">Fig. 19: The structure of knowledge database.</note><note type="content">Fig. 20: Illustration of the whole character matching “ ” with “ ”.</note><note type="content">Fig. 21: The first sample set contains 300 characters which were written (under constraint) by six writers to form 1800 samples.</note><note type="content">Fig. 22: The 32 radicals used in creating radical database.</note><note type="content">Fig. 23: The second sample set contains 156 characters.</note><note type="content">Table 1: </note><note type="content">Table 2: </note><note type="content">Table 3: </note><note type="content">Table 4: The tabulated results of radical extraction</note><note type="content">Table 5: The tabulated results of hierarchical radical matching scheme (for training set)</note><note type="content">Table 6: The tabulated results of hierarchical radical matching scheme (for testing set)</note><note type="content">Table 7: The recognition speed of hierarchical radical matching scheme</note><subject><genre>Keywords</genre><topic>Radical extraction</topic><topic>Character pattern detection</topic><topic>Radical matching</topic><topic>Knowledge matching</topic></subject><relatedItem type="host"><titleInfo><title>Pattern Recognition</title></titleInfo><titleInfo type="abbreviated"><title>PR</title></titleInfo><genre type="Journal">journal</genre><originInfo><dateIssued encoding="w3cdtf">200101</dateIssued></originInfo><identifier type="ISSN">0031-3203</identifier><identifier type="PII">S0031-3203(00)X0072-6</identifier><part><date>200101</date><detail type="volume"><number>34</number><caption>vol.</caption></detail><detail type="issue"><number>1</number><caption>no.</caption></detail><extent unit="issue pages"><start>1</start><end>185</end></extent><extent unit="pages"><start>15</start><end>35</end></extent></part></relatedItem><identifier type="istex">F56C0EB91A1089AD76635B08235F18333AE5E664</identifier><identifier type="DOI">10.1016/S0031-3203(99)00207-1</identifier><identifier type="PII">S0031-3203(99)00207-1</identifier><accessCondition type="use and reproduction" contentType="">© 2000Pattern Recognition Society</accessCondition><recordInfo><recordContentSource>ELSEVIER</recordContentSource><recordOrigin>Pattern Recognition Society, ©2000</recordOrigin></recordInfo></mods></metadata><enrichments><istex:catWosTEI uri="https://api.istex.fr/document/F56C0EB91A1089AD76635B08235F18333AE5E664/enrichments/catWos"><teiHeader><profileDesc><textClass><classCode scheme="WOS">COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE</classCode><classCode scheme="WOS">ENGINEERING, ELECTRICAL & ELECTRONIC</classCode></textClass></profileDesc></teiHeader></istex:catWosTEI></enrichments><serie></serie></istex></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/Istex/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000581 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Istex/Corpus/biblio.hfd -nk 000581 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Ticri/CIDE
|area= OcrV1
|flux= Istex
|étape= Corpus
|type= RBID
|clé= ISTEX:F56C0EB91A1089AD76635B08235F18333AE5E664
|texte= Optical recognition of handwritten Chinese characters by hierarchical radical matching method
}}
| This area was generated with Dilib version V0.6.32. |  |