Computer Vision 2
Semester: |
WS 2021 |
Type: |
Lecture |
Lecturer: |
|
Credits: |
6 |
Find a list of current courses on the Teaching page.
Lecture Organization
The following information is preliminary and may change at a later date due to new regulations in the pandemic. We hope for your understanding.
-
The lecture will take place in presence. For this, we have reserved lecture halls with sufficient capacity to comfortably sit the expected number of participants with an extended distance (at least 1 empty seat between participants in all directions). If the number of registrations further increases beyond that capacity, we have the option to move to a larger lecture hall.
-
All lectures will be recorded as screencast and will be made available to all lecture participants as a video in the moodle learning room (typically 1-2 days after the lecture due to the postprocessing effort involved). This way, nobody will have to miss a lecture slot due to potential illness or quarantine periods. If none of those conditions apply, we however strongly recommend attending the lecture in person, as it makes for a more wholesome learning experience.
-
For everybody's protection, lecture attendance will be subject to the usual "3G" conditions, i.e. physical attendance will only be possible for symptomless participants who can provide proof of either vaccination against, recovery from, or negative test of COVID infection in one of the formats accepted by the university. This proof will be checked upon entering the lecture hall, so please plan your arrival time accordingly.
-
We will also experiment with transmitting the lecture as a live zoom broadcast from the lecture hall in the spirit of a true hybrid format, but it is unclear whether bandwidth will be sufficient to support this.
Special regulations for students enrolled in the RoboSys program
RoboSys is a paid study program organized by the International Academy. Participation in RWTH lectures is made possible through individual lecture contracts with the teaching institutes. In the past years, the CV2 lecture was part of the RoboSys electoral curriculum, and corresponding lecture contracts have been initiated.
Starting with this semester, the RoboSys Steering Committee has decided no longer to include the CV2 lecture in the RoboSys curriculum and, consequently, has not concluded a lecture contract with our institute.
As a consequence, it will no longer be possible for RoboSys students to take the CV2 exam and get course credit for the course within RoboSys. The only exception to this policy are students who have already taken at least one failed try at the CV2 exam in previous semesters; they will still be able to participate in the exam in order to complete their studies. However, we will not be able to admit new RoboSys students to the CV2 exam. We're sorry for having to enforce this policy, but it hasn't been our choice. If you feel that this policy should change, please contact your RoboSys study coordinator. Nonetheless, we will grant you access to the lecture and learning material via moodle.
Lecture Description
The lecture will cover advanced topics in computer vision. A particular focus will be on state-of-the-art techniques for object detection, tracking, visual odometry and SLAM. There will be regular exercises accompanying the lecture.
Literature
In the last decades, Computer Vision has evolved into a rapidly growing field with research going into so many directions that no single book can cover them all. Some basic material can be found in the following books:
- Computer Vision - A Modern Approach, D. Forsyth, J. Ponce, Prentice Hall, 2002
- Multiple View Geometry, R. Hartley, A. Zisserman, 2nd edition, Cambridge University Press, 2003
- An Invitation to 3D Vision, Y. Ma, S. Soatto, J. Kosecka, S. Sastry, Springer, 2003
However, a good part of the material presented in this class is the result of very recent research, so it hasn't found its way into textbooks yet. Wherever research papers are necessary for a deeper understanding, we will make them available on this web page.
