Title :
A Morphological Approach to Persian Handwritten Text Line Segmentation
Author :
Shahraki, Abdollah Amirkhani ; Ghahnavieh, Amir Ebrahimi ; Mirmahdavi, Seyyed Abdollah
Author_Institution :
Dept. of Electr. Eng., Iran Univ. of Sci. & Technol., Tehran, Iran
Abstract :
This paper deals with extracting lines in handwritten Persian texts. In the presented method, at first, oblique lines and dots pertaining to words that are the main causes of two lines sticking to each other are eliminated. Then, using morphological operations, the remaining components of lines stick to one another. Finally, each line´s border is determined using contours. This method was tested on a database of the First Competition on Line Segmentation in Persian Hand- Written Documents, which was held in conjunction with the First Pattern Recognition and Image Analysis Conference (PRIA 2013) in Birjand, Iran. The results indicated proper accuracy of the proposed method.
Keywords :
handwritten character recognition; image segmentation; PRIA 2013; Pattern Recognition and Image Analysis Conference; Persian hand-written documents; Persian handwritten text line segmentation; handwritten Persian texts; line extraction; morphological approach; Accuracy; Databases; Educational institutions; Image analysis; Image segmentation; Morphological operations; Pattern recognition; text line segmentaion; morphology; Persian handwritten text;
Conference_Titel :
Computer Modelling and Simulation (UKSim), 2014 UKSim-AMSS 16th International Conference on
Print_ISBN :
978-1-4799-4923-6
DOI :
10.1109/UKSim.2014.93
