Project of the Scientific Grant Agency VEGA 2/0043/13

Project duration: 07/2011 - 10/2015

Principal investigator: prof. RNDr. Ing. Ivan Bajla, CSc.

The project is oriented towards the methods for complex time series analysis that are inspired by the theory of nonlinear dynamic systems, the research into biologically inspired hierarchical network models of pattern recognition, and the other complex network models. The main goals are: -optimization of the reconstruction of the dynamics of the underlying systems in multi-dimensional state space and design of novel algorithms of time series prediction and noise reduction, -extension of the biologically inspired network model of HTM (Hierarchical Temporal Memory), -study of interconnection of visual recognition with attention that will improve positional and scale invariances of pattern recognition. The methods will be applied to one- and multi-dimensional biosignals and to other experimental data with recognizable deterministic features, as well as to clustering and recognition of multi-dimensional time series and gray-level visual scenes with objects in clutter environment. 

Selected publications