Scientists have come up with a new system that counts shoppers by looking at their shoes. The shoes could provide a helpful marketing tool without the privacy invasion of analyzing CCTV images.
Stephan Richter and colleagues at the Hasso Plattner Institute in Potsdam, Germany, came up with the idea while working on a way to help software distinguish people gathered round large tabletop touchscreen computers.
The computer tends to struggle to keep track of who is doing what if people move around the screen while working.
To solve this, the team developed a system called Bootstrapper which works out who's standing where using a downward-facing Kinect camera to recognise their shoes.
By measuring how a shoe distorts the infrared grid projected by Kinect's depth camera, Bootstrapper acquires unique, high-resolution 3D images of each person's footwear.
The team found that the system has an accuracy of 96 per cent and will present their work at the Computer Human Interaction conference in Austin, Texas, in May.
The system could allow Kinect cameras mounted near to floor level to count the number of customers entering a shop, as well as determine which products they are most attracted to, said team member Patrick Baudisch, also at the Hasso Plattner Institute.
The idea has promise, said a spokeswoman at shopping behaviour analyst Synovate Retail Performance, based in Milton Keynes, UK.
"If Bootstrapper could differentiate by shoe size it could avoid the counting of children, who are not seen as retail opportunities," New Scientist quoted her as saying.