Oak Ridge National Laboratory (ORNL) is training next-generation cameras called dynamic vision sensors to interpret live information. Unlike a traditional digital camera that records large amounts of information in frames, a dynamic vision sensor, or DVS, transmits changes in light intensity on a per-pixel basis. Individual pixel locations are recorded and time-stamped to the microsecond, creating “data events” that are processed by a neuromorphic network. “Because the DVS records only changes in what it sees, there is no redundant data,” said Kemal Fidan, an ORNL intern. This capability enables the sensors to be fast, power-efficient, and effective in a wide range of light intensity. The DVS could be applied in robotics and could improve autonomous vehicle sensing. Dynamic vision sensors at Oak Ridge National Laboratory were trained to detect and recognize 11 different gestures, such as waving and clapping, in real time. The resulting image shows movement on the pixel level. This video shows a spinning coin. Courtesy of Kemal Fidan/Oak Ridge National Laboratory, U.S. Department of Energy.
