DocumentCode
476229
Title
A Size-independent method for Chinese character writing structure assessment
Author
Qi, Heng-Nian
Author_Institution
Sch. of Inf. Eng., Zhejiang Forestry Univ., Hangzhou
Volume
5
fYear
2008
fDate
12-15 July 2008
Firstpage
2754
Lastpage
2759
Abstract
On-line real-time assessment on Chinese character writing can be divided into three steps: stroke, stroke order and structure normative assessments. Through the first two steps the written Chinese characters can be regarded identifiable. Structure normativeness assessment can be identified as a measure of calligraphy, that is to assess whether the writing good or bad. The paper presents a size-unlimited Chinese character writing structure normative assessment method. According to the intrinsic structural characteristics of Chinese characters, three assessment features and their expressions such as symmetry, uniformity and compaction are designed. For the assessment feature generalization, the bound box is obtained from the track of the writing. Then support vector machine method is adopted to grade the writing level with differences of the assessment features between the written character and the corresponding standard template character. The experiment shows that the method can well reflect the quality of writing.
Keywords
support vector machines; text analysis; Chinese character writing structure assessment; online real-time assessment; size-independent method; size-unlimited normative assessment; stroke order; structure normative assessments; support vector machine method; written Chinese characters; Character recognition; Compaction; Cybernetics; Forestry; Handwriting recognition; Lattices; Machine learning; Support vector machines; Testing; Writing; Chinese character writing; Compaction; On-line assessment; Structure normativeness; Symmetry; Uniformity;
fLanguage
English
Publisher
ieee
Conference_Titel
Machine Learning and Cybernetics, 2008 International Conference on
Conference_Location
Kunming
Print_ISBN
978-1-4244-2095-7
Electronic_ISBN
978-1-4244-2096-4
Type
conf
DOI
10.1109/ICMLC.2008.4620875
Filename
4620875
Link To Document