Автоматизированная система выделения движущихся объектов на видеоизображении на базе рекуррентной нейронной сети

Abstract

This paper describes the recurrent and spiking neural network models used for detection of moving objects on video and creation of a motion detector. They released the moving object detection automated system based on the recurrent neural network and investigated ways of integration of programs developed using object-oriented programming languages (e.g. C#) with the visual processing framework ‘AForge.NET’ and the multimedia framework ‘DirectShow’. The most effective methods of the framework ‘AForge.NET’ are selected for visual processing