Search academic texts

Information × Registration Number 0214U007484, 0111U005378 , R & D reports Title Modelling visual attention on the base of winner-take-all artificial neural network popup.stage_title Head Tymoshchuk P. V., Registration Date 08-10-2014 Organization Lviv Polytechnic National University popup.description2 There are investigated existing mathematical models of visual attention. Itti model as one of the modern and widespread was selected to further improving. IKN-model was detailed analyzed, possible ways to improve are given. As a result of the replacement mechanism of finding focus - neural network on linear search algorithm has been made significantly improve visual attention model performance, which fundamentally allows its application to problems of machine vision. We propose to change RGB system of colors to HSV. The effectiveness of the functioning model of visual attention at different sizes of saliency map was investigated too, we saw that model gives better detection of objects both on artificial and natural images using larger saliency maps.5481 Product Description popup.authors Бренич Яна володимирівна Степанюк Сергій Іванович Тимощук Марта Павлівна Тимощук Павло Володимирович popup.nrat_date 2020-04-02 Close