Organizers: Qadeer Khan, Mariia Gladkova

Please forward any queries related to the lab to intellisys-ss22@vision.in.tum.de.

Assignment to the lab is done via the matching system. In addition to applying to the matching system please remember to also send your application documents latest by 16 February 2022 . The email format is presented during the preliminary meeting (see the slides).

The preliminary meeting took place on 4 February 2022. The slides are available here.

Note that the course goes by the name Lernbasierte Ansätze für autonome Fahrzeuge und intelligente Systeme on the matching system.

Learning-based approaches have recently made tremendous progress in the computer vision and robotics research community. In this practical course, we will solve challenging real-world problems in the area of self-driving cars and intelligent systems. Topics will be in the direction of visual localization and mapping, control, and semantic understanding. We will also explore the synergy of learning-based methods with classical geometry-based approaches such as visual odometry and 3D reconstruction. Some applications may also involve working on massive datasets including unstructured or unordered data.

• Good knowledge of the Python language and basic mathematics such as linear algebra, analysis, probability and numerics etc. is required.
• Good knowledge of a deep learning framework such as PyTorch, TensorFlow, etc.
• Participation in at least one of the offered deep learning lectures at TUM is required. For e.g. 1, 2, 3 etc.
• OR participation in at least one of the lectures / labs covering the basics of Multi-View Geometry. Some example courses include: 1, 2, 3, etc.
• Other courses with matching content may be considered. Please highlight this in your application.

Programming tasks will be given in the initial weeks to get the participants up to speed. Afterwards, students will work in groups of max. 2 persons on research oriented projects. To review progress and assist with resolving any issues, students are invited to meet the supervisors on a weekly basis.

Time: Tuesdays 11 am - 1 pm
Room: Online (details communicated to the participants)