AI Based Anomaly Detection for Bank Security Against Theft

A critical component of video surveillance research and real-world applications is the detection of anomalous events. In order to improve public safety, more and more surveillance cameras are being installed in public spaces like roadways, crosswalks, banks, and shopping centers. Identification of odd occurrences, such as car accidents, crimes, or illicit activity, is a crucial duty in video surveillance. Because anomalous events are less often than normal activities, the goal of an efficient anomaly detection system is to quickly identify departures from the average and identify the abnormality’s temporal span. One way to think of anomaly detection is as an early stage of video analysis that separates abnormalities from normal patterns. Using classification techniques, an anomaly can be further classified into particular activities once it has been discovered. An overview of anomaly detection with an emphasis on banking operations is given in this paper. Many every day and recurring transactions in the banking industry affect a number of stakeholders, including staff members, clients, debtors, and outside parties. Early detection of these anomalies can reduce or even eliminate any potential bad effects. In order to distinguish between normal and abnormal events, this work employs an anomaly detection technique based on machine learning.