21 October 1988
Vol 242, Issue 4877

About The Cover

Camera image overlayed with discontinuities in texture (yellow), motion (orange), and stereo depth (green). The discontinuities are computed from the output of early vision modules coupled with brightness gradient data using Markov random fields. The union of the discontinuities produces a “cartoon” which is used by a parallel recognition algorithm. See page 436. [T. Poggio et al., Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139]