Chinese newspaper layout analysis with antecedent compartmental lines

There exist many compartmental lines in Chinese newspapers, which make layout analysis very inconvenient. This paper proposes an approach based on compartmental lines, which extracts compartmental lines first, then segments the whole page with these lines, finally analyzes the interior area further using rectangle extension model. This method makes efficient use of the coordinate of the compartmental lines and avoids setting so many parameters in the components merging procedure, so it is more flexible. This method can locate compartmental lines accurately. Experiment shows that it is a page layout analysis algorithm with more universality.