Authors:
Zahra Gharaee
1
;
Peter Gärdenfors
1
and
Magnus Johnsson
2
Affiliations:
1
Lund University, Sweden
;
2
Lund University and NRNU MEPhI, Sweden
Keyword(s):
Self-organizing Maps, Neural Networks, Action Perception, Hierarchical Models.
Related
Ontology
Subjects/Areas/Topics:
Agents
;
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Cognitive Robotics
;
Computational Intelligence
;
Enterprise Information Systems
;
Health Engineering and Technology Applications
;
Human-Computer Interaction
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Methodologies and Methods
;
Neural Networks
;
Neurocomputing
;
Neurotechnology, Electronics and Informatics
;
Pattern Recognition
;
Physiological Computing Systems
;
Robotics and Automation
;
Self Organizing Systems
;
Sensor Networks
;
Signal Processing
;
Soft Computing
;
Theory and Methods
;
Vision and Perception
Abstract:
We present a novel action recognition system that is able to learn how to recognize and classify actions. Our system employs a three-layered neural network hierarchy consisting of two self-organizing maps together with a supervised neural network for labelling the actions. The system is equipped with a module that pre-processes the 3D input data before the first layer, and a module that transforms the activity elicited over time in the first layer SOM into an ordered vector representation before the second layer, thus achieving a time invariant representation. We have evaluated our system in an experiment consisting of ten different actions selected from a publicly available data set with encouraging result.