Please check the blackboard for up-to-date announcements!
Please do not forget to signup for the presentations using this spreadsheet.I know that you know most of the topics only by their title at this point but you will have some idea about their data after tomorrow’s lecture. So you can go check the datasets and see which problem’s data attracts you more 🙂
PS. We need at least one presenter for each week. If the topic you want is already taken, feel free to sign up as the second presenter but I might move you to another topic if there is any left without a presenter at the end of this week.
The first half of the course which concerns multiple view geometry contains some linear algebra. Therefore, it might be a good idea to review your linear algebra notes from your undergrad studies. Or you can watch the first two lectures of the Multiple View Geometry course by Daniel Cremers at TUM which are designed for the exact same purpose. Consider it as linear algebra for computer vision researchers 🙂
The rest of the course is also a great resource if you want to go further into detail on the topic at some point.
Here is a nice short explanation of camera intrinsics and extrinsics: video.
- Syllabus [pdf]
- Survey [pdf]
- TA: Sadra Safadoust (ssafadoust20[at]ku.edu.tr)
- Review Template (Overleaf)
Please copy and modify the file called “Spring2021/review_temp.tex”. You can use Overleaf or do it locally.
- Presentations (Spreadsheet)
- Introduction [PDF]
- Sensors, Datasets & Benchmarks [PDF]
- Stereo Estimation [PDF]
- Multi-view Reconstruction [PDF]
- Assignment (zip)
- Paper: Unsupervised Learning of Depth and Ego-Motion From Video, CVPR 2017 [PDF] [code:tf] [code:PyTorch]
T. Zhou, M. Brown, N. Snavely, and D. G. Lowe