Semantic Paint improving computer vision

Understanding what we see in the world around us might seem like a straightforward task to a person, but for computers that lack the context we have acquired over a lifetime of experience, it is anything but simple.

By allowing people to supply information about objects by physically touching them in the real world, this system gives us a quick and easy way of training computers to understand their local environments.

The ultimate aim is to use it to provide crucial information for augmented reality smart glasses that are being developed in Oxford to help people who are partially-sighted.

BBC Click finds out more.

More at BBC.com/Click and @BBCClick.