Signal processing approach for breath prediction pattern recognition

Michał Twardochleb, Tomasz Król

Download PDF

Citation:

M. Twardochleb, T. Król, "Signal processing approach for breath prediction pattern recognition", Journal of Theoretical and Applied Computer Science, vol. 7, no. 3, pp. 51-60, 2013.

Export to BibTex

Keywords:

pattern recognition, signal processing, neural network

Abstract:

In this paper, a new approach of signal processing for breath prediction pattern recognition is proposed and further analyses are presented. In order to extract key values from raw data, a shift from time domain to phase space has been utilized. It helped to achieve clearer peak-to-peak measurements which are crucial for breath prediction pattern recognition. Based on a special software tool for breath prediction pattern recognition several different algorithms have been compared. As a result, a reduction in error rate can be achieved when applying a new signal processing approach in comparison to the previous designs.