Title :
Hand-drawn line removal from Bangla printed document images
Author :
Tamali Banerjee;Samit Biswas
Author_Institution :
Department of Computer Science and Technology, Indian Institute of Engineering Science and Technology, Shibpur Howrah-711103, India
Abstract :
Every document holds some sorts of information. Paper documents especially old paper document provide a wealth of information about every aspect of the society and civilization. So these paper documents are being converted to electronic form for better storage, processing and analysis. The people, especially students while reading a book or documents have a tendency to mark important lines through the pen. To preserve these types of documents electronically, these lines need to be removed. This paper proposed a method to remove hand drawn lines from scanned paper document images. As there is no any benchmark database, we also proposed a dataset consisting of document images with hand-drawn lines (made it publicly available). The proposed method is tested on this dataset and the results are encouraging.
Keywords :
"Junctions","Conferences","Computer graphics","Information security","Optical character recognition software","Filling","Computer science"
Conference_Titel :
Computer Graphics, Vision and Information Security (CGVIS), 2015 IEEE International Conference on
DOI :
10.1109/CGVIS.2015.7449904