Symbolic/neural recognition of cursive amounts on bank cheques

Abstract
A symbolic/neural approach to recognize unconstrained handwritten cursive amounts in bank cheques is proposed. Features like ascenders and descenders are extracted from the binary image of the amount. Depending on the features extracted, some words are recognized entirely symbolically, some words entirely neurally, and the remaining both symbolically and neurally. Results of preliminary experiments are provided.

This publication has 4 references indexed in Scilit: