Abnormal behaviour detection in surveillance videos using post estimation method – A deep dive

People from all walks of life have long been deeply concerned about concerns related to public safety. The quantity of security cameras set up to watch over private and public areas has significantly expanded in recent years. The capacity to spot odd human behaviour in recordings has become more crucial as video detection technology has advanced. The identification of deviant human behaviour is crucial, especially in student groups. The majority of currently used algorithms for detecting abnormal human behaviour are designed to detect outdoor activity, and their inside detection capabilities might be improved. Modern classrooms are primarily outfitted with monitoring technology, and students spend the majority of their time indoors. This study provides a novel framework for the detection of aberrant behaviours in humans when they are indoors and focuses on the detection of abnormal behaviours in indoor humans. By predicting the placement of the subject’s important joints, we are able to identify anomalous human behaviour in this project. This technique is capable of estimating not only one but numerous individuals in the videos.