Unusual Activity Detection for Prior Appraisal against Crime: A Review

Abstract

Suspicious activities seriously endanger at public areas and personal security. There are millions of video surveillance systems used in public areas, such as streets, prisons, holy sites, airports, and supermarkets. Video surveillance cameras are not intelligent enough to recognize unusual activities even at real time. It is essential to investigate the detection and recognition of suspicious activities contents from surveillance video. It is required to recognized scamper situation at real time from video surveillance for quick and immediate management. The objective of this paper is to review various implemented systems and their flaws. Most of the system uses Gaussian filter for classifying object from a video or frame as per the gesture recognized. Few systems are based on background or foreground subtraction that works with two layers; first one is background and second one is foreground i.e. object. But these
systems are good enough for simple backgrounds or less crowded areas. These systems are simply recognizing some basic activities such as walking, sitting, running, waving hands, clapping and many more instead of classifying unusual activities. A system is required that should be intelligent enough for recognizing unusual activities from crowded area at real time.