Analysis of Objects Classification Approaches Using Vectors of Inflection Points

Autor:	Mykola Dacko, Andriy Kaniovskyi, Viktor Krylov, Ivan Kit, Diana Zahorodnia, Galina Shcherbakova, Anatoliy Sachenko, Pavlo Bykovyy
Rok vydání:	2019
Předmět:	Set (abstract data type) Identification (information) Inflection point Computer science business.industry Euclidean geometry Centroid Image processing Pattern recognition Class (philosophy) Artificial intelligence Object (computer science) business
Zdroj:	Advances in Intelligent Systems and Computing ISBN: 9783030264734 ISDMCI
DOI:	10.1007/978-3-030-26474-1_11
Popis:	An increase of the automated video surveillance system operational performance was proposed by reducing the amount of data using the contour inflection points for object identification. The authors developed a structural statistic method to identify facial images. It allowed obtaining a set of vectors that consist of inflection points of every hierarchical layer of the processed object. The research described in this paper shows the comparison of such vectors identifying similarities and referring them to a particular class. Three objects classification approaches were proposed and analyzed: approaches based on centroid, sum and maximum. Two alternative schemes for the calculation of distance between inflection point vectors were also proposed and analyzed: the classical Euclidean meter and the modified one, based on distances from the coordinate origins.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::ca0d4a196dbab5826adfac8a08436d87 https://doi.org/10.1007/978-3-030-26474-1_11 Zobrazit plný text záznamu