Line Reconstruction and Segmentation of Words and Characters using Measures of Central Tendency and Measures of Dispersion

Abstract

This research concentrates on reconstruction of the output line segments of the paper in [1]. The Line Segmenting module of [1] in some scenarios segments the alphabets and the associated matras of a line text in two separate line segments. These line segments are reconstructed using the Reconstruct Module to produce a line text with all its alphabets and its associated matras. This module uses one of the measures of dispersions, that is, standard deviation to accomplish reconstruction of output line segments. Then words are segmented from the line segments using WordSegmenting Module. This module uses one of the measures of central tendencies, i.e. mean and one of the measure of dispersions i.e, standard deviation to achieve word segmentation. Then characters are segmented from words using CharacterSegmenting Module.