Abstract
Text embedded in video frames provides useful information for semantic indexing and browsing system. In this paper, we propose an efficient approach for automatic Arabic-text detection which combines edge information and maximally stable extremal region (MSER) method in order to extract text region candidates. These regions are, then, grouped and filtered on the basis of geometric properties such as area and orientation. Besides, we introduce a new geometric descriptor of Arabic text called baseline to improve the filtering process. Our proposed approach was tested on a large collection of Arabic TV news and experimental results have been satisfying.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
