During my postdoc at Caltech, I worked on analog VLSI vision chips that can compute the optical flow of an image focused with a lens directly onto the chip. These chips can be used as a real-time testbed for investigating biological motion integration algorithms, which can solve the aperture problem over their receptive field and respond meaningfully to transparent motion. It is also possible to do image segmentation through motion, and to integrate the optical flow over the whole chip to compute egomotion information.
For my Ph.D. research, I worked on information theoretic extraction of conjunctive rules from databases, applied both to discrete classification and function approximation. The function approximator was used to learn a nonlinear control system.In the ancient days, I worked on a novel hardware routing system (US patent number 5,181,017) for a message-passing parallel computer.