Title :
Segment and Semi Circle Based Geometric Characters of Bangla and An Efficient Relevant Method of Character Recognition
Author :
Abdullah-Al-Mamun, Khondaker ; Siddika, Nafisa Khanam ; Al Ameen, Mahmudul Faisal ; Sarker, Farhana ; Akbar, Md Mostofa
Author_Institution :
Dept. of Comput. Sci. & Eng., Ahsanullah Univ. of Sci. & Technol., Dhaka
Abstract :
This paper presents a new segment and semi circle based model (S-SCBM) to write Bangla vowel characters. All vowel characters are formulated by using arc(s), line(s) and segment(s). Here an efficient method for segment and semi circle based character recognition has been implemented. For character recognition, required features are number of line(s), arc(s) and segment(s). The feature vector consists of presence or absence of line(s), arc(s) and segment(s). In the feature vector, 0 indicates absence and 1 indicates presence of line(s), arc(s), and segment(s). From the method we get different index values for distinct characters. By using the index value any Bangla vowel characters can be recognized. Since only degrees of arcs, line and segment are considered, the method of character recognition is size independent
Keywords :
natural language processing; optical character recognition; Bangla vowel; character recognition; index value; segment based geometric characters; semi circle based geometric characters; Character recognition; Computer science; Optical character recognition software; Pattern recognition; Speech recognition; Text recognition; Writing;
Conference_Titel :
Electrical and Computer Engineering, 2006. ICECE '06. International Conference on
Conference_Location :
Dhaka
Print_ISBN :
98432-3814-1
DOI :
10.1109/ICECE.2006.355323