Aaron Hertzmann is both a computer scientist and a painter. Among his art-related programming projects, he has developed a way for computers to refashion images and animations to reflect different painting styles. Other work includes a stretched “canvas” that projects a continually-changing painting of the viewer and a learning-by-example system that animates characters based on motion-capture data.
Go to: http://mrl.nyu.edu/projects/npr/painterly/,
http://www.dgp.toronto.edu/~hertzman/art.html,
and http://www.mrl.nyu.edu/publications/style-machines/.
