Monday, March 28, 2005
I was reminded of the method of 'Chernoff Faces'; in Clifford Pickover's 1990 book: "Computers, Pattern, Chaos and Beauty" (p. 47-) .. I was led there by a pointer to some Java implementations of Chernoff faces. For those that have not seen this, its the use of cartoon faces, whose attributes (eyes, ears, mouth, hair, etc) are varied in their size and orientation based on some numerical measure. Specifically using a cartoon face is said to allow us to use our own facial pattern recognition ability to better understand complex multidimensional data. Can anyone point me to other implementations or examples of a similar idea ... linking numeric measures to the size of other aspects of an image in general? Is this idea used very often today?