Data-Set for Normal and Suspicious Behavior in ATM Booths

This research paper presents a Data Set for an innovative neural networking model designed to enhance the security of Automated Teller Machines (ATM) through the detection of suspicious activities in real-time video streams. The proposed model utilizes advanced computer vision techniques and deep learning methodologies to analyze video feeds from ATM cameras. By identifying unusual behaviors such as bending down and looking away from the screen, the system aims to provide robust intrusion detection capabilities, thereby preventing potential fraudulent activities. The model's effectiveness is evaluated through extensive experimentation, demonstrating its potential to enhance the security landscape of ATM operations significantly.