Date: 26 Oct 2001

On-Line Hand-Drawn Symbol Recognition Based on Primitives Separation and Fuzzy Inference

* Final gross prices may vary according to local VAT.

Get Access

Abstract

In this paper, a universal recognition method for hand-drawn symbols is presented which based on primitives separation and fuzzy inference. This technique first separates input figure symbol into primitives using a single-band integral algorithm, then a fuzzy inference framework is used to combine these primitives into complicate symbols. The features used in the inference process are very simple, so the figure set can be easily extended. The method was tested under regular input circumstance with an average recognition rate of 90 percent reported.