Using Fringe Maps for Text Line Segmentation in Printed or Handwritten Document Images

Abstract

Accurate segmentation of text lines from printed or handwritten documents is an important task in any document processing system. This becomes a challenging and complex problem due to several reasons. Situations arise when the text from neighboring lines overlaps the white space area, or touches text of the current line. Complications may also arise when due to varying skew, text lines curve along the page in varied trajectories. These situations are beyond the scope of common algorithms developed for some printed or handwritten documents. In this paper, we propose a novel approach based on fringe maps to generate segmenting paths between adjacent text lines. First we generate a fringe map for the input binary image, next we compute peak fringe numbers (PFN) to locate potential regions to find a separating path. PFNs between lines are used to generate a segmenting path to separate adjacent lines. The method is demonstrated on various types of examples including those with Indic scripts, both printed and handwritten.