Event Location: NSH 1507
Bio: Uwe Franke received the Ph.D. degree in electrical engineering from the Technical University of Aachen, Germany, in 1988 for his work on content based image coding.

Since 1989 he is with Daimler Research & Development working on vision based driver assistance systems. He developed Daimler’s lane departure warning system (“Spurassistent”) introduced 2000 on trucks. His special interest is in image understanding in complex situations as they occur e.g. at intersections. He has been working on real-time stereo vision since 1996. Recent work is on optimal fusion of stereo and motion, called 6D-Vision and scene flow.

Since 2000 he is head of Daimler’s Image Understanding Group and concentrates on vision for increased traffic safety. His team is specialized in object recognition and tracking based on stereoscopic special-temporal perception. Continuously working in the field of intelligent vehicles since 22 years, he is one of the most experienced experts all over the world. In 2002, he was program chair of the IEEE Intelligent Vehicles Conference in Versailles, France.

Abstract: The performance of future driver assistance systems depends on precision, robustness and completeness of their environment perception. The urban scenario in particular poses high demands on the sensors, since dangerous situations have to be recognized quickly and with high confidence. The dream of a car perceiving its environment with human like performance in order to realize accident free driving can only be reached if the car has two eyes working in stereo – that is my firm belief.

The talk will present the state-of-the-art in space-time computer vision. This covers real-time dense stereo analysis (running on a FPGA) as well as dense optical flow analysis (running on a GPU). Most known stereo systems concentrate on single image pairs. This prohibits the recognition of moving objects like pedestrians, if they are close to other dominant obstacles or partially hidden. A smart fusion of stereo vision and motion analysis is the key to overcome this deficiency. Two strategies have been developed. The so called 6D-Vision principle tracks points with depth known from stereo over two and more consecutive frames and fuses the spatial and temporal information using Kalman filters. Secondly, a scene flow variant will be described that runs at 5 Hz on a standard GPU. This approach simultaneously estimates depth and 3D-motion for every pixel of the image.

The high-quality spatio-temporal information is successfully used to model the world and to detect and track moving obstacles from the moving car. Pedestrian recognition significantly benefits from the dense stereo and motion information. For oncoming vehicles, besides speed and acceleration even the turn-rate can be determined reliably.

Modern schemes like GraphCut together with careful error propagation allow segmenting moving objects precisely even at large distances, where the depth uncertainties cannot be ignored. Efficient scene labelling becomes possible that decides for every pixel whether it belongs to the free-space, a moving object with known motion, a moving object with unknown motion or to the background.

Real-world experiments illustrate the high performance available in the experimental car – hopefully paving the way towards safer driving.