Odds and quads

These artworks were created by The Painting Fool, a program that "aspires to be an artist". The software was developed by Simon Colton, Michel Valstar and Maja Pantic, computer scientists at Imperial College London.

Graphic software such as Photoshop can create what looks like an Impressionist painting from a photograph almost instantly. The Painting Fool, by contrast, attempts to reproduce the process of painting. It does so by filming an object, splitting the image into regions of colour and simulating how a person might paint each area. It also imitates the effects of media such as chalk and acrylic.

Dr Pantic and Dr Valstar developed a system to enable the program to recognise emotions by scanning the facial muscles captured by images.

It has honed its skills, like generations of art students, by attempting to copy the Old Masters. It also produced these "emotionally enhanced portraits" of Meera Senthilingam, producer of the BBC radio programme The Naked Scientists.

The results, which can be seen online, won a British Computer Society Machine Intelligence Award in 2007.

Send suggestions for this series on the sector's treasures, oddities and curiosities to: matthew.reisz@tsleducation.com.

Already registered?

Sign in now if you are already registered or a current subscriber. Or subscribe for unrestricted access to our digital editions and iPad and iPhone app.

Register to continue  

You've enjoyed reading five THE articles this month. Register now to get five more, or subscribe for unrestricted access.

Most Commented

  • Elly Walton illustration (16 July 2015)

Whether in jest or not, sexist language shows an insensitivity to gender issues at odds with academic values, argues Dorothy Bishop

  • Tony Little, Eton College headmaster, 2007

Tony Little points to ‘increasing gap’ between teaching standards at sixth form and university

  • Tourists in rubber rings and flippers ready for snorkeling class

Dress to impress if you want students in your corner, claims US study

  • gold on scales

£246 million is big money but it is probably much less than the hit the research budget would take if the REF did not exist, says Paul Jump