Handwritten Arabic Text Recognition System using Window Based Moment Invariant Method

Main Article Content

Mowaffak Othman Al_Barraq
S.C. Mehrotra


A sliding window approach has been used for segmentation of the handwritten Arabic texts into atomic characters or sub characters
using features based on moment invariant technique. Each separated characters is represented by the n-dimensional space where n is number of
windows related to the character. The Recognition rate based on Euclidian distance and cosine () similarity approaches, is found to be 92.86


Keywords: Windowing; Handwritten Arabic Texts; Moment Invariants; Features Extraction; Curve Fitting and Representation.


Download data is not yet available.

Article Details