Figure-ground segregation using an analog VLSI chip

Our working, analog VLSI vision chip labels all points inside a given contour with one voltage and all remaining points outside this contour with another voltage. Its behavior is very robust, since small breaks in the contour are automatically "sealed," providing for figure-ground segregation in a noisy environment. This circuit with its networks of resistors and switches represents a step toward object-level processing, since a single voltage value encodes the property of an ensemble of pixels.