CAN鈥橳 find a face in the crowd? Not to worry, a human recognition system can spot people for you 鈥� even when their faces aren鈥檛 visible. Designed for , it recognises people by the clothes they are wearing. Their name is then overlaid on the headset鈥檚 video.
The system, called InSight, is part-funded by Google and was unveiled at the in Jekyll Island, Georgia, last week. It aims to help users find their friends and be spotted themselves in busy places like shopping centres, sports stadia and airports. Face recognition systems cannot be used for this, says InSight developer at the University of South Carolina in Columbia, because it is unlikely someone in a crowd will be looking straight at a headset鈥檚 camera.
So Nelakuditi joined forces with and colleagues at Duke University in Durham, North Carolina, to develop a recognition system based on a 鈥渇ashion fingerprint鈥� of a person鈥檚 outfit, from their clothes to their jewellery, badges and glasses.
Advertisement
This fingerprint is constructed by a smartphone app which snaps a series of photos of the user as they read web pages, emails or tweets. It then creates a file 鈥� called a spatiogram 鈥� that captures the spatial distribution of colours, textures and patterns (vertical or horizontal stripes, say) of the clothes they are wearing. This combination of colour, texture and pattern analysis makes someone easier to identify at odd viewing angles or over long distances.
Usefully, in terms of protecting people鈥檚 privacy, the fingerprint changes every time you change your clothes, so you can be anonymous again whenever you wish.
鈥淎 person鈥檚 visual fingerprint is only temporary, say for a day or an evening,鈥� says Nelakuditi.
In early tests using 15 volunteers, the team identified people 93 per cent of the time, even when they had their backs to the headset user. Matching data from the phone鈥檚 motion sensor with motion in the Glass field of view will boost accuracy.
鈥淭here are a lot of personal characteristics that make us unique,鈥� says , a biometrics specialist at the University of Southampton, UK. 鈥淐lothing and movement are highly related to gait 鈥� and gait has been shown to be unique.鈥�
The system could be used by someone who wants to attract attention to themselves and their CV at a job fair, or outside a stadium where they are selling a spare ticket, the team says.
It could perhaps even help people with a condition known as face blindness 鈥� a neurological disorder that makes it impossible to recognise others 鈥� by telling them the names of friends nearby.
This article appeared in print under the headline 鈥淵ou are what you wear when you鈥檙e tagged in a crowd鈥�
