Seminar on Current Works in Computer Vision

Prof. Thomas Brox

Computer Vision is a very active research field with many practical applications, for instance for industrial inspection, autonomous driving, robotics, or smart homes. The ultimate goal of Computer Vision is to imitate the great capabilies of the human visual system, allowing the computer not only to record images but also to interpret them. Research has made significant progress in recent years also due to deep learning.
In this seminar we will take a detailed look at the most interesting recent works. The focus will be on deep learning for computer vision and visual control problems. For each paper there will be one person, who performs a detailed investigation of a research paper and its background and 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.

Seminar: (2 SWS) Wednesday, 10-12am (hct), Room: 52-2-17 Contact person: Özgün Çiçek Beginning: Wednesday, April 26, 2017, 10:15 Allocation of seminar topics ECTS Credits: 4 Recommended semester: 6 (Bachelor), any (Master) Requirements: Background in computer vision or deep learning Remarks: This course is offered to both Bachelor and Master students. The language of this course is English. All presentations must be given in English. Topics will be assigned in the first meeting. If you do not show up in the first meeting, you will not have a place. If there are more interested students than places, places will be assigned by a mixture of motivation in the first meeting and priority suggestions of the system (in case it works). In particular, we want to avoid that people grab a topic and then jump off during the semester. Please have a coarse look at all available papers to make an informed decision in the first meeting. The listed papers are not yet sorted by the date of presentation. There is a related Blockseminar on Deep Learning for Biomedical Image Analysis offered by apl Prof. Olaf Ronneberger from Google DeepMind. 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.

---

Slides of first session with instructions for a good presentation
Powerpoint template (optional)

Papers:

Date	Paper	Questions	Presenting student	Slides	Advisor
21.06	Object instance segmentation	Questions	Silvio Galesso	Slides	Özgün Çiçek
28.06	Activity understanding	Questions	Thomas Leyh	Slides	Mohammadreza Zolfaghari
05.07	Depth map fusion with an octree convolutional network	Questions	Lukas Voegtle	Slides	Huizhong Zhou
12.07	Uncertainty estimation in deep networks	Questions	Patryk Chrabaszcz	Slides	Eddy Ilg
19.07	Visual imitation learning	Questions	Salih Hasan Siddiqi	Slides	Artemij Amiranashvili
26.07	Learning 3D Human Pose Estimation	Questions	Ikrima Bin Saeed	Slides	Christian Zimmermann