Disclosure and Sniff out of Moving Entity in Real World

Abstract

Entity disclosure & sniff out is being studied from years together and is one of the area where research is constantly carried out. In today’s world it is a great challenge to generate an approach which is robust, accurate & high performing. How an entity is disclosed & sniffed out is defined as one of difficult task. One of the visual features, say a particular color is used as representation of an entity then the method discloses as a entity all the pixels with same color. On another side it is very hard to disclose accurately the face of particular person with full details (different actions & lightning changes) and to recognize, track. The biggest challenge is tracking entity in a video, since the entities are in motion. If a camera is fixed at appropriate point, as the entity moves in the area covered by it there is dramatic change in the entity image. This change occurs from three sources: if there is any change in the target posture, lightning changes and Due to change in camera setup property it is not possible either partially or fully to see what we wished to see. The videos that are captured under various environment needs to be understood in order to know the activities of entity, the task is very challenging and is used by many applications for companies, scientific research, educational institutions. What motivated in studying this problem is to create a system where moving entity surveillance in real time can be disclosed and sniffed out.