Text Line Extraction from Complex Layout Documents
Mohan Singh Brij, Ankush Mittal, Vivek Chand, Debashish Ghosh · International Journal of Computer Applications · 2011
are numerous stylish documents which do not have the traditional text layouts where printed text regions are not parallel to each other. Such complex layouts make text line extraction challenging due to multi-orientation of paragraphs. This paper introduces a system for the text line extraction from the complex layout documents. Proposed method is based on the concept of dilation and histogram profiling. The text regions are extracted using dilation and food fill based approach, then paragraph orientation is determined and individual text lines are extracted. The accuracy of extracted text lines are evaluated using the new proposed concept that is also based on the histogram profiling. The results of proposed approach on the complex layouts are promising.