Automated Scene Flow Data Generation for Training and Verification

Automated Scene Flow Data Generation for Training and Verification
Oliver Wasenmüller, René Schuster, Didier Stricker, Karl Leiss, Jürger Pfister, Oleksandra Ganus, Julian Tatsch, Artem Savkin, Nikolas Brasch
ACM Computer Science in Cars Symposium (CSCS) ACM Computer Science in Cars Symposium (CSCS-2018), Munich, Germany

Abstract:
Scene flow describes the 3D position as well as the 3D motion of each pixel in an image. Such algorithms are the basis for many state-of-the-art autonomous or automated driving functions. For verification and training large amounts of ground truth data is required, which is not available for real data. In this paper, we demonstrate a technology to create synthetic data with dense and precise scene flow ground truth.