Self-Reporting System for ATM Incidents Detection Using Machine Learning Techniques

ATM security incidents are a significant concern for financial institutions, leading to financial losses and service downtimes for users. A Self-Reporting System for ATM Incident Detection (SRSAID) is proposed to address these issues using tampering images and R-CNN fine-tuned by a regression model for automatic detection. Python 3 is employed on Raspberry Pi, utilizing GPS and GSM modules for periodic updates. Experiments reveal a 96% accuracy with ssdlite_mobilenet_V2, outperforming ALEXNET's 80%. A web-based interface supports decision-making, reducing downtime experiences.