Friday, September 28, 2012

Automated Art Appreciator

An experiment, published in a recent issue of ACM Journal on Computing and Cultural Heritage, on using algorithms to have machines analyze visual clues from 1,000 paintings of 34 well-known artists, and grouping them into relevant genres, proved that the computers perform better than untrained humans in quantifying the subtle differences between closely related schools of art.

No comments:

Post a Comment