Seminar for computer science, games engineering and computational science and engineering (CSE) master students (IN2107).

Reconstruction of 3D geometry has always been a central application in Computer Vision research. This seminar will give an overview of the broad range of methods that can be applied to retrieve geometric information from sensor data. The focus will be on methods that are able to reconstruct geometry with a high level of detail. The approaches we look at range from shape-from-shading and photometric stereo to surface reconstruction from point cloud data. We will cover both classical approaches, such as variational methods or differential equations, and newer methods from the field of Deep Learning.

• Organizers: Erik Bylow, Christiane Sommer
• Email: 3dgeometry-ws19@vision.in.tum.de (please address all requests to this address only!)
• If you are accepted to the seminar in the matching system, you can expect to get more information from us by mid-August. Please do not contact us before!

All information as well as the slides from the pre-meeting can be found here (password protected).

• Day: Wednesday, every second week
• Time: 10:00 - 12:00
• Start: October 30th
• Location: 02.09.023
• Each session there will be two talks.
• All talks are held in English.