Project Lab Deep Learning in Computer Vision
Winter Semester 2016/2017
In this project lab groups of students will work on selected topics in deep learning (deep neural networks) for problems in computer vision. This includes the practical implementation with modern deep learning frameworks. Results will be presented in a talk at the end of the lab. Concrete topics follow the current state of the art and change from term to term.
Students will have access to dedicated lab machines, including a workstation with 4 nVidia GeForce Titan X cards.
Quick links
Organization
| Module no. | 20-00-0980-pp |
| Event type | project laboratory (PP6, 9CP, 6SWS), elective |
| Being taught | regularly, usually every winter semester |
| Organizational meeting | 18.10.2017, participation mandatory |
| Location | see TUCaN |
| Enrollment | max. 18 participants |
| Lecturer | Prof. Stefan Roth, Ph.D. |
Content
- Familiarization with the current state-of-the-art in deep learning methods for computer vision
- Familiarization with state-of-the-art software frameworks for deep learning
- Implementation and, if applicable, extension of a deep learning method in computer vision; done in teams
- Presentation of the final result
Learning goals
Through their successful participation, students acquire in-depth knowledge on deep neural networks and their applications in computer vision. They are able to analyze, modify, and apply state-of-the-art techniques in this area. Moreover, they practice their abilities for presenting their results and for collaboration in teams.
Degree courses
Can be taken for credit toward BSc / MSc Informatik, MSc Visual Computing, and MSc Autonome Systeme. BSc / MSc Students from other departments and study programs, e.g. Computational Engineering, Mathematik, Elektrotechnik, IST, or Physik are welcome, though academic credit may need to be arranged.
Prerequisites
- Solid programming skills in C/C++ or Python or Lua
- Prior or concurrent registration for Computer Vision I
Access restricted section: Log in to see this section.