Classification of Background Subtracted Videos Using Neural Network-Learning Classifier

In this project we present the concept of effective classification for background subtracted videos by using learning classifier-feed forward neural network with back propagation to conquer the open problem in the context of the complex scenarios.eg:while picturing the videos in some application like cloudy (or) misty areas the object in video will be less clarity with naked eye even after the background subtraction also. In context of existing system the back ground subtraction is to apply the video brick extraction at any location of the scene, which is in the avi format. Next is to find the threshold value of bricks by using spatial-temporal information of the video based on the intensity. The threshold values separate the foreground and background of the bricks by applying background subtraction. It will get rid of the background of the each and every frame and finally run as video. Here we proposed the effective classification for the output video that has underwent by background subtraction & reveal the type of the object more precisely then existing system by using learning classifier–feed forward neural network with back propagation. Keywordsbackground modeling, spatial-temporal representation and feed forward neural network-learning classifier.