Automated real-time risk assessment for airport passengers using a deep learning architecture

A Holistic Review of Cybersecurity and Reliability Perspectives in Smart Airports

“…Trolley RFID locator [109], video surveillance systems [110], electronically controlled doors [111].…”

“…For example, RFID Transmitters for the detection of luggage trolleys [109] in parking lots or inside the airport is such an application, as often travelers will not return the trolleys they used to its proper location. Further applications include, the video surveillance systems [110] and electronically controlled doors [111] that monitor the passengers in order to swiftly detect suspicious behavior and controlling access to certain restricted and possibly high-risk areas of the airport.…”

“…The identified faces can then be cross-referenced with law enforcement databases that maintain images of wanted terrorists and criminals, informing authorities in the case of a positive match [169], [170]. Another application is behaviour and intention analysis, where DL models are trained to scan video streams and detect irregular or suspicious movements of individuals, unusual poses, facial expressions and concealed weaponry [110], [171]- [174]. Such systems can be deployed in critical infrastructure like airports, where irregular movement of a passenger could signify a potential threat.…”

Automated Real-time Anomaly Detection in Human Trajectories using Sequence to Sequence Networks

“…In our approach we assume that the trajectories are readily available, namely we can track each object's exact po-sition for a specific time interval (similar to [28]). In the general case, trajectories can efficiently encode the information regarding motion in the video.…”