Trilingual Script Separation of Handwritten Postal Document

Postal automation is a topic of research over the last few years. There are many works towards the postal automation in USA, UK, Japan and Australia, but for Indian postal automation there is a few significant works. This paper deals with tri-lingual word-wise handwritten script identification for Indian postal automation. In the proposed scheme using Run Length Smoothing Algorithm, postal document is segmented into lines and then into words. Using fractal-based features, busy-zone based features and topological features, a neural network classifier is used for word-wise Bangla, English and Devnagari scripts identification. Overall accuracy of the proposed system is at present 96.79%.