Lost luggage causes chaos at train stations and airports. It's not only a problem for the person who's misplaced their bag, but also sparks disruption to station staff and security personnel, and delays other travellers when a platform or terminal is closed because of safety concerns.
Solving such problems is the central idea behind social innovation: finding solutions not for one person or organisation, but for all of society. We tolerate CCTV, for example, even though it invades our personal privacy, because of the broader benefit to our security and safety.
And while train stations and airports are usually bristling with CCTV cameras, scrolling through video takes time and plenty of eyes — both of which are in short supply when an abandoned backpack is spotted. "Manually looking for a person through video takes a lot of manpower," says Peter Jones, head of marketing for digital solutions at Hitachi. "It's whole teams of officers looking." There's a better way: letting artificial intelligence do the hard work.
Hitachi's AI tool takes a standard camera feed, and analyses it in real-time using a neural network, looking for full-body characteristics. Footage can be live or previously taken, with the system tracking individuals as they move through a public space not via facial recognition, but through more than 100 physical attributes, such as height, clothing, or what they're carrying. Such a system has a wide range of social-innovation use cases, helping everyone from parents to police. Looking for a lost child? Staff can search for lone children wearing specific clothing colours. Tracking a suspicious individual for police? Enter in the eye-witness description, and the technology will narrow down suspects, letting staff select the right target and then track them through their journey, even when the characteristics that initially identified them are no longer visible.
"AI can be taught to recognise anything," says Jones. The system can differentiate 120 categories from head to toe, including hair length and shade, whether they have a bag, even their shoe colour. But none of those characteristics are personally identifying, which helps protect privacy – another social innovation making an existing public-safety good such as CCTV even better for all of us. That means that as well as tracking unknown people, the system can follow the path of each commuter or shopper for marketing or planning, without involving personally identifying data, such as their name, season ticket number or face. Alongside avoiding issues raised by using such private data, the characteristic-based tracking was found to be three times more accurate than facial recognition.
By having that privacy protection built in, the smart AI video system can be used beyond public safety, for a range of social goods. A construction site can use it to keep watch over workers, simply by searching for anyone wearing a bright yellow vest. Marathons can use it to track runners without any other identifying details, and spot anyone leaving the sidelines and joining the runners, making it easier to see if they're a threat, a steward or a prankster. Shopping malls can use it for marketing and data analytics, tracking shoppers through their buying spree. Police sifting through hours of retrospective footage for a suspect can track them from the scene of the crime both forwards and backwards in time. "There's a lot of uses, from marketing and crowd control to looking for anomalies such as a bag left alone," Jones says. "We can use AI in a variety of ways."
Once the system knows what it's looking for, it can hunt out anyone matching those attributes, getting better at narrowing down the options with every successful find as it learns more detail about what the individual was wearing. And if the criminal dumps their jacket or chucks on a hat? Security staff can still search on their other attributes — such as height or shoes — in order to pick up the digital scent. "You'll get your answer much more quickly," says Jones.
And don't forget that left-behind rucksack. Security staff could scroll back to when the bag was left behind, and then ask the system to track the owner. If the person immediately sprinted from the station, it's time to evacuate. If they're standing on the station concourse asking staff for help, the lost luggage and traveller can be easily reunited. "For public safety, it has a lot of applications," says Jones. "There's been several tests in airports of people leaving their things behind. It could reduce the chances of a terminal being closed."
But the AI-based video analysis is only one piece of the puzzle for Hitachi, with a wide range of visualisation tools for public safety and smart city technologies available through its social innovation work. "We offer the pieces to make a bigger solution, so customers take whatever they're interested in," says Jones. "It's not one solution, but a range of tools to make a system."
Hitachi Visualisation Suite pulls together AI analysis and third-party CCTV camera networks, sensors such as microphones that listen for gunshots, licence-plate recognition platforms, and social media, aggregating the data into a single view for analysis and offering full situational awareness for faster responses. Such tools are scalable and layered, so public authorities such as policing agencies can pick and choose the tools they need to assemble the public safety system that's right for them — and for all of us. Now that's social innovation.
For more, listen to WIRED's special social innovation podcast at the top of this article.
--
Modern life is saturated with data and technologies are emerging nearly every day – but how can we use these innovations to make a real difference to the world?
Hitachi believes that social innovation should underpin everything it does, so it can find ways to tackle the biggest issues we face today.
Visit Social-Innovation.Hitachi to learn how social innovation is helping Hitachi drive change across the globe.
This article was originally published by WIRED UK
