Call For Paper Volume:5 Issue:8 Aug'2018 |

Detection and Classification of Interesting Events Using ANN Algorithm

Publication Date : 26/02/2016

Author(s) :

Jisha Shajan , Prema Mani.

Volume/Issue :
Volume 3
Issue 2
(02 - 2016)

Abstract :

The analysis of motion and behaviors in crowded environment using a surveillance system constitutes an issue for public security. The proposed system focuses on detecting and classifying interesting events in crowded video. A newly introduced method based on swarm theory, histogram of oriented swarms is used to capture the motion of crowded scenes, together with the histogram of oriented gradients are used to extract the appearance characteristics of a video sequence. Both motion and appearance combined to form final descriptor, which effectively characterize each scene and detect anomaly. Because of the infinite number of anomalies that is derived from each scene, it is difficult to provide examples of all possible abnormal classes, so an Artificial Neural Network (ANN) classifier is chosen.ANN aiming to identify any irregularities deviating from the normal pattern. This leads to achieves higher accuracy and classifier capable of detecting different kinds of anomalies at a low computational cost.

No. of Downloads :



Web Design MymensinghPremium WordPress ThemesWeb Development

Detection and Classification of Interesting Events Using ANN Algorithm

February 11, 2016