image copyright Evolv Technology caption of the image Peter George hopes to solve the paradox of protection
This very high alarm rate prompted Evolv to combine a AI software with radar to reduce false alarms and keep crowds flocking to a place without annoying interruptions.
It's not just about looking for the shape of a gun like as defined by the software, but also for small l shards packed in a confined space to create o shardsbus around an explosive device, as was tragically demonstrated in the Manchester Arena bombing in 2017.
Evolv software engineers have written algorithms that interpret shapes like signatures, with the outlines of knives and pistols cataloged as reasons to alert an operator.
Metal density is a key indicator of the presence of a weapon, and Mr. George says his colleagues programmed the system to respond in 1 / 38th of a second.
The AI software analyzes the results it gets from each crowd and these are automatically uploaded into Evolv's own database.
Each ark sold is integrated to the personal archives of Evolv via the Internet. It is therefore updated from other scanners. The system evolves as the user base grows, or as Mr. George says, "we can make everyone smarter when we share thes data we collect.
Evolv hopes to add thermal imaging capabilities to its scanners during the year 2020. In theory, this could detect people with high temperatures, the 'one of the symptoms of Covid-19.
Dr Simon Worrell is an immunologist specializing in pandemic planning who works for the health and safety consultancy Collinson. He says using AI to assess human scans "sounds good, but I don't know how effective it will be ".
Taking paracetamol, for example example, will lower the temperature of a carrier of Covid-19. And basic precautions remain paramount. "Physical distancing is important.
Image copyright image caption Experts say temperature controls are not enough to them alone to detect Covid-19
He thinks that adhering to layers of protection, temperature controls, masks and staying away, is the way to defeat the virus.
Canadian threat protection specialist Patriot One Technologies asked its AI team to research how its existing systems, designed to identify a person carrying a gun in a crowd, can be set to capture people whose temperature is high.
But Martin Cronin, managing director of Patriot One and a former diplomat at the British Foreign Office, calls for caution. "It 's not that easy to detect temperatures. You can't just point a thermal imager at a crowd. Factors like ambient temperature or if someone has a high temperature due to exercise can affect the accuracy of scanning devices. People should be wary of bold claims to have temperature analysis answers. "
Part of the challenge is tuning the AI software to ignore these distractions and isolate indicators that are important in keeping virus carriers away from crowded places. Patriot One wants to pinpoint a series of anomalies, including the lack of approved face masks, which mean a visitor deserves more attention.
It was quite difficult to train the systems to search for concealed weapons.
Patriot One found that the radar did a great job watching under coats and jackets to find what he thought were armes. Unfortunately, the number of false alarms was unreasonably high.
Image copyright Patriot One image caption The Patriot One system can be disguised under plant pots
The AI was designed to help, in the form of machine learning technology that could sort out the shapes captured by the radar scan and assess if they really posed a threat.
With a helping hand en With AI the success rate increased, but it had not yet reached a stage where it could be unleashede on the public without the common objects being confused with the tools of a terrorist.
The system only worked well in some environments, while in others the signals were clogged with outside elements such as broadcasts from consumer electronics.
More enterprise technology
The answer was even more technology. Patriot One bought several companies, with additional expertise in AI and detection.
Video footage of guns and knives were introduced into the software of AI, teaching it to discriminate between different objects.
And the company has connected its surveillance technology to security cameras suspended in familiar transparent bulges of hallway ceilings at customer sites.
Mr. Cronin put the ingredients of AI together, pooling data from different systems. He likens it to how humans combinet all their senses to understand the world around them.
image copyright image caption AI can monitor information from multiple sources
He claims AI software has now eliminated interference from building noise and electronic signals, removing false positives that blocked early trials. And it presents the technology as leaving individuals anonymous. "We're looking for things, not people.
But the algorithms that underpin smart systems canbe rigged by mundane events.
At a high school in the United States, a gun-detecting scanner was puzzled by the sudden appearance of umbrellas on a heavy rainy day. Recycled to identify an umbrella as a non-threatening object, the software could then be updated on the Patriot One user network.
Different societies have different problems. In North America, the priority is to detect firearms. For potential UK users, the system can search for knives. But in both cases, the idea is to install a discrete security architecture.
So the next time you walk through a corporate reception area or the entrance to a public building, take a look at these large planters that frame the door as you enter. You may have been checked for weapons. Or maybe, one day, a virus.