Computer Vision [CV]
This lecture discusses important algorithms from the field of computer vision. The emphasis lies on 3-D vision algorithms, covering the geometric foundations of computer vision, and central algorithms such as stereo vision, structure from motion, optical flow, and 3-D multiview reconstruction. Participants of this advanced course are expected to bring experience from prior lectures either from the field of pattern recognition or from the field of computer graphics.
Dates & Rooms:
Wednesday, 10:00 - 12:00; Room: 01.255-128
- READ ME FIRST: We have to limit the lecture to 20 participants. Thank you for the large interest in the class - we are more than full. Please do not send any additional emails for registration. Places will be distributed today (Sunday); everyone who sent an email in the past days will receive a reply today.
- The lecture is on Wednesday, 10am, room 01.255-128 (Cauerstr. 11).
Exercises are offered on Monday, 4pm and Tuesday, 12pm, both in Huber-Cip (room 0.01-142).
- To pass this class, it is necessary to pass the practical exercises, and to pass an oral exam.
- If you wonder whether this lecture is for you, consider browsing the textbook a little bit.
The lecture follows the beautiful book by Szeliski: "Computer Vision: Algorithms and Applications", which is also available online. In the lecture, we will quickly go over the introductory chapters, and then talk about chapters 6, 7, 8, 11, 12 (and maybe 14, if time suffices).
- Exercises will be in C++ using the open-source computer vision library openCV. If you feel like your C++ should be slightly freshened up, please do so.
(Quick self-assessment: what means "const"? What is the difference between a reference and a pointer? What is a smart pointer? How to declare/define classes? How and where should I de-allocate memory that has once been allocated?)
Note that we will not need more advanced language concepts like templates.