Almost all have seen, security guards at airport or big shopping malls, scanning human body for suspicious object. The guard are checking manually. A sci-fi film “Total Recall” (1990) showcased hi-tech scanners. Medium to large scale companies are in a race of building innovative products and applications using artificial intelligence. Recently, a Boston based startup, Evolv Technology, has developed artificial intelligence based high-speed body scanners. The company will conduct its first public trials at Union Station in Washington DC, in Los Angeles’s Union Station metro. This article highlights features and issues of Evolv’s AI Powered Body Scanner.
AI Powered Body Scanner – Features
Interesting feature of the body scanner is that People can walk through a security scanning gate without slowing down. Security guard can see ok sign or a photo showing suspicious areas highlighted. Computer vision and machine learning based new device can scan human body within fraction of second, to spot guns and bombs. The system can scan 800 people an hour. Also there is no need to remove their coins , keys, or cellphones.
Solid state micro-antennas within AI Powered Body Scanner body scanner directs radar beams over person walking through the gate. It pick up the reflections back. The body scanner system has been trained to spot distinctive patterns of all kinds of objects, such as firearms, gun, knives, etc. The scanner is equipped with facial recognition capability to recognise person.
The company claims that the AI Powered Body Scanner is faster and cheaper than existing millimetre-wave scanners. Therefore the new device could bring airport security to higher level. Looking at recent large-scale attacks at transit facilities, we need to have a system that can screen large number of peoples and detect weapons and explosives effectively and efficiently.
Testing and Issues
Elkins, an AI professor at San Diego State University, who is developing his own AI security technology, said that public tests will be very much useful. Testing high-performing detection algorithm in lab may not reflect true real world issues. The system will be test while people carrying all the things they would normally carry.
One of the issue in AI Powered Body Scanner is false positives. People may be incorrectly flagged as carrying something suspicious. If too many people are falsely flagged, then organisations will really feel, not to use it. One of the problem with algorithms, as opposed to human, is that an attacker may reverse-engineer a machine can fool the system.
Countermeasures may be possible, but the research community does not yet have a good sense of how to construct them. But, in fact such systems will definitely become reality because of the fast technological development and demanding applications.