YOLO Algorithm-Based Suspicious Activity Detection in ATM Surveillance

Utilizing the YOLOv3 algorithm for detecting suspicious activities in ATMs presents an efficient approach to bolstering ATM security. Renowned for its real-time object detection capabilities, YOLOv3 accurately identifies various objects in ATM surveillance footage, including individuals, vehicles, and other pertinent elements. The method employs YOLOv3 to monitor and detect objects of interest within the surveillance footage. Detected objects undergo scrutiny to identify suspicious behavior, such as attempts to obscure the keypad or card reader, prolonged stationary positioning, or loitering near the ATM. This proposed approach offers a reliable and proactive means of identifying suspicious activities in ATM surveillance, promptly alerting security personnel as necessary. The effectiveness of this method can be assessed using performance metrics such as detection accuracy, precision, recall, and F1 score, further enhanced through tailored fine-tuning of the YOLOv3 algorithm on specific ATM surveillance datasets.