Inhalte

This research group explores non-rigid 3D registration under topological changes—a frontier challenge in computer vision! This semester-long journey will dive into aligning 3D models to real-world images of objects, even when these objects deform, tear, or fracture. With applications in medical imaging, robotics, and augmented reality, this group is perfect for those passionate about computer vision, geometry or optimization. This research group is tailored towards providing hands-on experience with advanced algorithms, exposure to real-world problems, and even the chance to contribute to a potential research publication. The semester will be split between lectures, intensive study of existing research and short research assignments that gradually aims to push the boundaries of the state-of-the-art on this topic. Pre-requisites are: a) fundamentals of computer vision, especially the basics of multi-view geometry, b) fundamentals of non-convex optimization methods (familliarity with convex optimization is helpful, but not strictly necessary), and c) good programming skills either in MATLAB or Python.

Kontakt

sengupta@zib.de

Link zum Vorlesungsverzeichnis

folgt