Selection of informing aggregates of diagnostic parameters in tasks of image recognition

1Pavlenko, VD, 1Fomin, AA
1Odesa National Polytehnic University, Odesa, Ukraine
Kosm. nauka tehnol. 2001, 7 ;(Supplement1):043-047
https://doi.org/10.15407/knit2001.01s.043
Publication Language: Russian
Abstract: 
The methods of leading multiclass images recognition tasks to sequence of recognition tasks with two classes are analyzed. The method of selection informing aggregate of diagnostic parameters in multiclass images recognition tasks, based on optimization of a goal function is offered. The effectiveness of different types of goal functions application is explored.
References: 
1. Zagoruiko N. G., Elkina V. N., Emel'yanov S. V., and Lbov G. S. OTEKS Applied Software (for Data Analysis), 160 p. (Finansy i Statistika, Moscow, 1986) [in Russian].
2. Tou J., Gonzalez R. Pattern Recognition Principles, Transl. from Eng., 411 p. (Mir, Moscow, 1978) [in Russian].