Search academic texts

Information × Registration Number 0213U003422, 0110U002636 , R & D reports Title Models and methods of multidimentional data granulation and interpretation popup.stage_title Head Mashtalir V.P., Registration Date 05-02-2013 Organization Kharkiv National University of Radioelectronics popup.description2 Сontext clusterization videodata methods and models; micro-situation data analysis methods and models; images normalization methods and construction invariant to geometrical transformations signs methods with the use image projections are worked out. A study and estimation efficiency of different approaches and artificial perception methods are undertaken in the complex analysis task of the graphic stages. A methodological base is chosen, namely semantic going near the images analysis. The semantic going near artificial perception is considerably anymore flexible, than traditional statistical, and assumes the construction of invariant IUS, that does not depend on a subject domain. These advantages matter very much for the graphic analysis stages, as a subject domain of initial images can not be limit. Generalized going near using of likeness measures for thr tasks of multimedia data analysis solved. Also the hybrid methods of statistically-semantic analysis and recognition of the graphic stages, that does not depend on a subject domain, and allow to execute treatment of graphic information real-time, were worked out as a result. The algorithm of recognition and analysis of the stages is worked out based on semantic neuron network. Product Description popup.authors Білоус І.А. Білоус Н.В. Братчун А.С. Власенко Н.В. Горбач Т.В. Гороховатський В.О. Гороховатський О.В. Кіношенко Д.К. Кобилін О.А. Красов А.И. Кузьменко Є.Ю. Куліков Ю.В. Ляшенко В.В. Мантула Є.В. Машталір С.В. Передрій О.О. Полякова Т.В. Путятін Є.П. Сакало Є.С. Удовіченко А.Г. Шляхов В.В. Щербінін К.С. popup.nrat_date 2020-04-02 Close