Comparative analysis of various feature descriptors for efficient ATM surveillance framework
Vishal Sanserwal, Vikas Tripathi, Zhiqian Chen, Monika Pandey · 2017
Automatic Teller Machine (ATM) is a beneficial service of a bank to carry out banking transactions in public space. The ATM allows the access to the bank account for cash withdrawal, check balance or transfer money. However, besides facilitating the banking needs, the ATM's lack in providing the security against the several ATM fraud like money snatching and attacks on customers. ATM has gain importance worldwide to enhance security and protection of individual and infrastructure. Feature descriptors play an important role in vision based surveillance in order to extract relevant features from video sequence for activity classification. In this paper we present comparative analysis of different descriptors like Gradient based descriptor i.e. HOG(Histogram of Oriented Gradients) and shape based descriptors like Hu Moments and Zernike Moments for effective abnormal activity detection in a single view point scenario like ATM premises.