Unusual Activity Detection in Crowd using Deep Learning

Modern technology is making people’s life easier, but the safety of life is also the most important thing. Crowded places like public events, stadiums, festival grounds, rally affects not only the comfort level of humans, but it also increases the risk of safety of pedestrians and other civilians. Heavy crowds may lead to major accidents, crowd-crush and causing an overall control loss. To reduce the risk of civilians in heavy crowds, we have worked with technology to manage the crowd. The applications of crowd management are manifold, ranging from crowd counting to human computer interaction. Research articulated here is focused on how to detect any unusual event in crowds at the early stage using modern technology like Deep learning so that it can be handled and managed timely and causes least harm to civilians. The concept of Convolutional Neural Network is utilized for processing of images and videos.