The criminal offences in the ATM kiosk are happening very commonly in recent days. A fully automated ATM surveillance system is the need of present era intended for detecting suspicious actions in the surveillance system and to trigger the proactive steps before the incident to occur. An innovative methodology proposed in this paper, which deals an automation of video surveillance in ATM kiosk and detect any type of potential criminal activities. In this system, an innovative methodology is proposed for automated ATM surveillance System using skeleton-based action recognition neural networks and IoT sensors. Multiple layers of detection techniques used to confirm the activity as suspicious. Skeleton-based action recognition by part-aware graph convolutional networks is used for detecting suspicious human action using the NTU RGB-D data set. Aadhar enabled finger print scanner which is integrated with ATM is used to fetch the demographic information from aadhar server. IoT proximity sensor is used to recognize any trial to block the vision of surveillance camera. Similarly, any physical attack made on ATM will be identified using IoT pressure/gas sensors. Suspicious sound generating during the criminal offence is also considered to confirm the activity as suspicious. Once, the activity is confirmed as suspicious, demographic information of the suspect will be fetched from aadhar server maintain by unique identification authority of India (UIDAI) and initiate the proactive steps and warning procedures.

ATM, part-aware graph, convolutional networks , NTU RGB-D data set, Surveillance System

[1] S.Shriram, S. B.Shetty, V.P. Hegde , KCR Nisha, Dharmambal.V, “Smart ATM Surveillance System”, 2016 International Conference on Circuit, Power and Computing Technologies [ICCPCT].
[2] M. Baranitharan, R. Nagarajan, G. ChandraPraba, “Automatic Human Detection in Surveillance Camera to Avoid Theft Activities in ATM Centre using Artificial Intelligence”, International Journal of Engineering Research & Technology (IJERT)-NCICCT – 2018. Volume 6, Issue 03.
[3] K. Archana, P. B. Reddy , A. Govardhan, “To Enhance the Security for ATM with the help of Sensor and Controllers”, International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS-2017).
[4] S. Aakur, D.Sawyer,S.Sarkar, “Fine-grained Action Detection in Untrimmed Surveillance Videos”, IEEE winter applications of computer vision workshops Computer Science and Engineering, University of South Florida, Tampa, 2019, pp-38-40.
[5] K. Soomro, H. Idrees, M. Shah, “Online Localization and Prediction of Actions and Interactions”, IEEE Transactions on pattern analysis and machine intelligence.
[6] O.Ayankemi ONI , "A Framework for Verifying the Authenticity of Banknote on the Automated Teller Machine (ATM) Using Possibilistic C-Means Algorithm", International Journal of Scientific Research in Computer Science and Engineering, Vol.6, Issue.2, pp.57-63, 2018.
[7] V.K. Jain, N. Tripathi, "Speech Features Analysis and Biometric Person Identification in Multilingual Environment", International Journal of Scientific Research in Network Security and Communication, Vol.6, Issue.1, pp.7-11, 2018
[8] Y. Qin, L. Mo1, C. Li1, J. Luo1, “Skeleton-based action recognition by part-aware graph convolutional networks”, Springer-Verlag GmbH Germany, part of Springer Nature 2019.
[9] X. Suna P. WuaSteven, C.H.Hoi, “Face detection using deep learning: An improved faster RCNN approach”, Neurocomputing (2018).