A New Technique for Segmentation of Handwritten Numerical Strings of Bangla Language
Md Aktaruzzaman, Md. Farukuzzaman Khan, Ahsan-Ul-Ambia · International Journal of Information Technology and Computer Science · 2013
Segmentation of handwritten input into individual characters is a crucial step in connected handwriting recognition systems.In this paper we propose a robust scheme to segment handwritten Bangla numbers (nu merical strings) against the variab ility involved in the writ ing style of different ind ividuals.The segmentation of digits from a nu mber is usually very tricky, as the d igits in a Bangla number are seldom vertically separable.We have introduced the concept of Degenerated Lower Chain (DLC) for this purpose.The DLC method was proved efficient in case of segmenting handwriting d igits in our experiments.Ten pages of handwritten Bangla numerical strings containing 2000 individual dig its that construct 700 numbers written by five d ifferent writers of variable ages were segmented by the developed system.The system achieves more than 90% segmentation accuracy on average.