Multi-agent System for Recognition of Hand Postures

  • Authors:
  • Mariusz Flasiński;Janusz Jurek;Szymon Myśliński

  • Affiliations:
  • IT Systems Department, Jagiellonian University, Cracow, Poland 31-110;IT Systems Department, Jagiellonian University, Cracow, Poland 31-110;IT Systems Department, Jagiellonian University, Cracow, Poland 31-110

  • Venue:
  • ICCS 2009 Proceedings of the 9th International Conference on Computational Science
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

A multi-agent system for a recognition of hand postures of the Polish Sign Language is presented in the paper. The system is based on a syntactic pattern recognition approach, namely on parsable ETPL(k) graph grammars. An occurrence of a variety of styles of performing hand postures requires an introduction of many grammar productions that differ each from other slightly. This makes a construction of a grammar within a parsable class ETPL(k) dubious. Dividing a whole grammar into sub-grammars and distributing them to agents allows one to solve the problem.