Egocentric Computer Vision


15 October: Please note that the seminar will be held for the first time on Monday, 27 October, i.e. the second week of the semester.


This research seminar will explore the challenges of egocentric computer vision and wearable head-mounted vision devices, such as Google Glass. First, we start with a brief introduction and literature review on existing work in this area. The main part of the course consists of a project in which students will develop, implement and evaluate an example application. The organisers will assist participants in shaping and formulating project ideas, which will then be presented and discussed within the group. Based on the discussions and feedback these ideas will be prototyped and evaluated on datasets or in a user study, as appropriate. The seminar will conclude with a final presentation of the project outcomes and a project report.


The research seminar is intended for Master students in Computer Science, Visual Computing and Media Informatics. Very good programming skills (C, C++, Python or Java) are required as a substantial part of the seminar will consist of implementing a prototype application, which will be graded as part of the project work.
Successful completion of at least one of the following core lectures is required: Image Processing and Computer Vision, High-level Computer Vision/Graphical Models, Machine Learning, Human-Computer Interaction, or Computer Graphics.


Dr. Yusuke Sugano
Dr. Andreas Bulling

date and time

every Monday from 10am – 12am in room 633, E1.4 (MPI-INF)


Winter term 2014


To register for the seminar please use this link. Registrations by e-mail will not be considered.

link to university course overview