Event-based Image Processing using a Neuromorphic Vision Sensor

A DVS (Dynamic Vision Sensor) is a device that generates spikes just like a human retina does. The spikes are created in the pixels by illumination changes. A software project called jAER is a way of communication with the DVS; jAER is developed in Java, uses OpenGL to render the image and drivers to communicate with the DVS through USB. The Processing in jAER must be frame-free (event-based). This paper presents a DVS and jAER analysis, and describes the method to do new applications in jAER, as well as presents two applications done in jAER: one is the tracking of objects and the other is the construction of the image from the spikes.