An analog VLSI velocity sensor system for depth perception

The authors present a neuromorphic analog very large scale integration integrated circuit (IC)-based system for determining the relative depth of objects in a scene. This system called the depth through motion parallax (DTMP) system computes the depth of an object in its field of view using the DTMP algorithm - an algorithm that is based on psychophysical studies of vision in humans. The IC computes the apparent velocity of the object from the measured transit times (the time it takes for the image to move from one pixel to the next) and consumes an average power of less than 2 mW. To compute the depth of a stationary or moving object, the authors accelerate the observer using a motion controller and obtain the apparent acceleration (change in the apparent velocity) rather than just the apparent velocity. The experimental results demonstrate that the sensor can be used to provide accurate relative depth information of the objects based on their apparent accelerations. The description of a fully self-contained system based on the prototype DTMP system is also presented.