Mahidol University Logo
Faculty of ICT, Mahidol University
 

Admissions

Printable Version

 

THE RECOGNITION OF THAI HANDWRITTEN-CHARACTERS USING FEATURE-BASED APPROACH

 

TITLE THE RECOGNITION OF THAI HANDWRITTEN-CHARACTERS USING FEATURE-BASED APPROACH.
AUTHOR SURASIT KIWPRASOPSAK
DEGREE MASTER OF SCIENCE PROGRAMME IN COMPUTER SCIENCE
FACULTY FACULTY OF SCIENCE
ADVISOR JARERNSRI L. MITRPANONT
CO-ADVISOR SUKANYA PHONGSUPHAP
 
ABSTRACT
There are a very limited number of researches that are effective enough to extract most of the significant features of Thai handwritten characters. Many of them have concentrated on the extraction of only a few features mainly for Thai optical character recognition systems. This research aims to develop a set of concrete feature extraction algorithms to be used in the recognition of off-line Thai handwritten characters by using the feature-based approach. These algorithms are used to exploit inherent characteristics or prominent features of Thai characters. Both the characteristics and the common human writing behaviors of Thai characters were studied and analyzed. The decision trees were used to classify Thai characters that share some common features into five classes. A set of key features of Thai characters was identified. The major features covered were an end-point (EP), a turning point (TP), a loop (LP), a zigzag (ZZ), a closed top (CT), a closed bottom (CB), and a number of legs. These features were defined as standard feature or the “Thai Character Feature Space.” Then, we defined the 5x3 standard regions used in mapping the standard features of all Thai characters. The result of the mapping process was the “Thai Character Solution Space,” which can be used as a fundamental toolfor recognition.The twelve algorithms used to determine each feature of Thai characterswere designed and developed along with the recognition system. This includedthe algorithms to determine an incomplete loop and a filled loop. Eachalgorithm has been tested thoroughly by using of more than 44,600 Thaicharacters handwritten by 22 individuals from 100 documents.The experimental results of each algorithm are summarized as well asthe performance. The feature extraction rate is as high as 98.66% with theaverage of 93.08% while the recognition rate is as high as 99.19% with theaverage of 91.42%. The results indicate that ourproposed algorithms are wellestablished and effective.
KEYWORD THAI CHARACTER RECOGNITION / THAI HANDWRITTEN CHARACTER RECOGNITION /FEATURE EXTRACTION / FEATURE EXTRACTIONALGORITHM / FEATURE-BASED APPROACH

 

Go to Top

 

ICT Building, Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhonpathom 73170 Tel. +66 02 441-0909 Fax. +66 02 849-6099
Mahidol University Computing Center, The Faculty of ICT, Mahidol University , Rama 6 Road, Rajathevi, Bangkok 10400 Tel. +66 02 354-4333 Fax. +66 02 354-7333