Low-Latency Visual Odometry with the Dynamic and Active-pixel Vision Sensor (DAVIS)

I am excited to share our first results on Low-latency, Event-based 6-DoF Visual Odometry with the DAVIS sensor (which combines a standard camera and an event camera in the same pixel array). Our algorithm uses an event-based adaption of SVO on event-based feature tracks: features are first detected in the grayscale frames and then tracked asynchronously in the blind time between frames using the stream of events.The features are then fed to an event-based visual odometry algorithm (adapted from SVO) that tightly interleaves robust pose optimization and probabilistic mapping. We show successfully tracking of the 6-DOF motion of the sensor in natural scenes. This is the first work on event-based visual odometry with the DAVIS sensor using feature tracks.The algorithm runs in real time. More info in our IROS'16 paper Low-Latency Visual Odometry using Event-based Feature Tracks and EBCCSP'16 paper: Feature Detection and Tracking with the Dynamic and Active-pixel Vision Sensor (DAVIS). If you want to know more about the DAVIS: the original paper is "A 240x180 130dB 3us latency global shutter spatiotemporal vision sensor", while the most recent development, which involves an RGB VGA resolution sensor is "Design of an RGBW color VGA rolling and global shutter dynamic and active-pixel vision sensor".