Seminar on Current Works in Computer Vision

Prof. Thomas Brox

Computer Vision is a very active research field with many practical applications, for instance in quality control, robotics, or driver assistance systems. The goal of Computer Vision is to imitate the flexibility and robustness of the human visual system. Research has made significant progress in recent years particularly due to deep learning. Most relevant research in Computer Vision has shifted to improving deep learning methods.
In this seminar we will take a detailed look at some of the most interesting recent works. For each paper there will be one person, who performs a more detailed investigation of a research paper and its background and who will give a presentation. The presentation is followed by a discussion with all participants about the merits and limitations of the respective paper. You will learn to read and understand contemporary research papers, to give a good oral presentation, to ask questions, and to openly discuss a research problem.

Due to the Corona crisis, the seminar will be entirely held as an online seminar via Zoom, i.e., the student presentations will be given using screen sharing and the discussion will be online as well.

Seminar: (2 SWS) Thursday, 10:15-11:45, Online via Zoom (Room 52-2-17 if university reopens) Contact person: Özgün Çiçek Beginning: Watch the following lectures before May 16: About the seminar Giving a good presentation Proper scientific behavior If you want to participate, register in HisInOne for the course and send an email with your name and your paper priorities (S1-S7, favorite paper first) to Özgün Çiçek before May 16. There is no online meeting planned. If you have questions, simply write an email to Özgün Çiçek or Prof. Brox. Recommended semester: 6 (Bachelor), any (Master) Requirements: Background in computer vision Remarks: The language in this course is English. There is a related Blockseminar on Biomedical Image Analysis offered by apl Prof. Olaf Ronneberger from Google DeepMind. The seminar will be held as an online seminar via Zoom. The link is valid for all sessions. You need a password to join. This was sent to you by email together with additional information, such as joining by phone in case your internet is not working. The presenter should join the session 15 minutes earlier to make sure they are familiar with the equipment and can share their screen. This only works when the host is present and allows these functions, so you need not test it at other times: it won't work. Important! Please get in contact with your advisor as soon as possible, and at least 4 weeks before your presentation Submit your presentation outline to your advisor at least 2 weeks before your presentation and meet with your advisor. Submit your presentation slides to your advisor at least 1 week before your presentation and meet again. All participants must read all papers and answer a few questions. The questions will be available in the 'Questions' column of the table below at least one week before the corresponding presentation. The answers must be sent to the advisor of the paper before the paper is presented. All participants must attend all online sessions.

---

Slides of the introductory lecture
Powerpoint template for your presentation (optional)

Papers:

Date	Paper	Questions	Presenting student	Slides	Advisor
25.06	Positional encoding in networks	Questions	Leonhard Sommer	Slides	Huizhong Zhou
02.07	Video representation and reasoning	Questions	Hakan Sarp Aydemir	Slides	Philipp Schroeppel
09.07	Network training	Questions	Anupam Kakkar	Slides	Yassine Marrakchi
16.07	Self-supervised representaton learning	Questions	Saurav Shanu	Slides	Svenja Melbaum
23.07	Large-scale semi-supervised learning	Questions	Sergio Izquierdo	Slides	Anton Boehm
30.07	Large-scale transfer learning	Questions	Guilherme Miotto	Slides	Sudhanshu Mittal
30.07	Learning visual respresentation from instruction videos	Questions	Julia Guerrero	Slides	Maria A. Bravo