Algorithm learns to understand natural beauty

Researchers at the University of Warwick have trained an algorithm to rate scenic images from one to 10.

To do this, they focused the algorithm on 200,000 images rated by people on the website ScenicOrNot, which has images of scenes that cover 95% of Great Britain.

The algorithm was able to recognise the difference between countryside landscapes and pictures of derelict buildings and rate them accordingly with the most beautiful scenes scoring ten.

The researchers say that their algorithm could be useful as a tool to aid planning and preservation of beauty spots.

See more at Click's website and @BBCClick, external.