 |
 |
 |
 |
 |
 |
Abstract
This course will provide a clear and in-depth tutorial on state-of-the-art techniques in camera motion estimation and 3D reconstruction from image sequences. Often this problem is referred to as Structure-from-Motion (SfM), Structure-and-Motion (SaM), or Visual Simultaneous Location and Mapping (Visual SLAM). The course starts with a brief introduction to projective geometry and different mathematical camera models. Next, linear entities are presented that described the geometry of multiple views, e.g., the fundamental matrix or the trifocal tensor. Using these foundations, different automatic approaches for camera motion estimation from images or video will be described. Furthermore, interactive and automatic techniques for 3D reconstruction of the static 3D scene geometry are introduced. The course also covers extensions to the basic algorithms, e.g., how to handle multiple moving cameras that capture the same scene simultaneously or how to reconstruct moving objects.Short Biography
Thorsten Thormählen studied Electrical Engineering (Dipl.-Ing.) at the University of Duisburg and received a PhD (Dr.-Ing.) from the faculty of Electrical Engineering and Computer Science, University of Hannover, Germany. From 2005 till 2007, he worked as a postdoctoral research fellow at the School of Computer Science, University of Adelaide, Australia. Currently, he is the head of the independent research group "Closed-Loop Structure-and-Motion" within the Max Planck Center for Visual Computing and Communication, Saarbrücken, Germany. His research interests are computer vision and computer graphics, especially, camera motion estimation and 3D reconstruction from video. He has been working on a number of national and international projects, including the well known "Voodoo Camera Tracker" and "Videotrace".Back to the program page