A Transputer-Based System for Multi-Stroke Character Recognition

Abstract

A multi-stroke handwritten character recognition system is presented in this paper. A suitable set of Fourier Descriptors is used as features set to describe each stroke. A man-machine interactive algorithm was utilized to accomplish the learning phase and to recognize unknown samples. A PC based prototype was realized at first to test the system. In order to speed up the recognition process, the design and a first prototype of a Transputer-based system have been realized too. Some preliminary experimental results obtained applying these machines to handwritten numerals are reported in the paper.