Abstract:

This paper describes a method to recognize sign language by Petri nets. A word in sign language is composed of a series of specific primitive signs. The primitive signs, e. g. standing palm, closing fist, are detected by template matching. The occurrence of a primitive sign stands for a transition in Petri nets. We propose a prototype Petri net to track signs. A testing word in sign language is recognized by checking the increase of the tokens in the  target places of the Petri net.