Research on Pattern Recognition Applied for Volume Estimation of Blood Chamber with Matrix of Optical Sensors

P. Marczyński, G. Konieczny and T. Pustelny Department of Optoelectronics, Silesian University of Technology, Akademicka 2, 44-100 Gliwice, Poland

Full Text PDF

The following article describes research on possibility of using pattern recognition algorithms in the optical measurement system for estimation of the blood chamber volume in the Polish Ventricular Assist Device (POLVAD). The optical system is being developed at the Department of Optoelectronics, Silesian University of Technology, Poland. Data analysis methods include a feature subset selection algorithm involving principal components analysis and objective function as quality criterion. The analysis takes into account 17 patterns reflecting particular volumes. The k-nearest neighbours method is used as pattern classifier. The pattern recognition system was initially designed for an array of gas sensors and this paper describes its further development.

DOI: 10.12693/APhysPolA.124.498 PACS numbers: 42.30.Sy, 47.54.-r, 43.60.Lq