Suspicious Activity Detection Using Deep Learning

Abstract

Big data applications are currently taking up majority about attention & space in business & academia. Unstructured big data largely comes from surveillance videos. This paper's main goal is towards provide a concise introduction towards video analysis utilising deep learning techniques in order towards identify suspicious activity. Our primary interest is in use about deep learning algorithms towards identify number about participants, number about involved individuals, & activity occurring in a crowd under all circumstances. We are able towards attain security thanks towards video analysis. Security can be described in a variety about ways, including identifying theft, spotting violence, etc. Simply put, practise about identifying unusual (abnormal) human activity is known as suspicious human activity detection. towards do this, we must break down video into individual frames, & processing these frames enables us towards examine people & their behaviour. In this system, there are two modules: an activity detection module & an object detection module.The presence or absence about an object is determined by object detection module. following module will determine whether activity is suspicious or not after detecting item.