64-439 Integrated Seminar Cognitive Computer Vision

Instructors: Prof. Dr. Simone Frintrop; Dr. Mikko Lauri

Event type: Seminar

Displayed in timetable as: CCV - Int.Sem

Hours per week: 2

Credits: 3,0

Language of instruction: English

Min. | Max. participants: - | 40

Comments/contents:
Seminar including practical programming exercises associated with the lecture Cognitive Computer Vision. Topic are closely related to the lecture content, e.g., covering topics such as visual attention modeling, saliency computation, and object discovery.

Learning objectives:
Getting hands-on experience with implementing methods in cognitive computer vision, e.g.  saliency computation.
Critically analysing and evaluating scientific literature and giving research presentations.

Didactic concept:
Course language is English.
One part of the course will consist of practical programming exercise, the second part of a seminar.

The programming exercise will be done in C++ with the OpenCV library. Knowledge of an object oriented programming language is mandatory, but not necessarily with C++. We will have a brief introduction into C++ and OpenCV at the beginning of the course, but additional self-study is required in case of no C++ knowledge. The programming exercise will lead step by step to building one complete saliency system that is able to locate salient regions in images.

For the seminar part, each student can select one research paper out of a provided paper pool, present this work to the other students and write a short report about this work.

The course requires pre-knowledge of basic computer vision concepts, such as convolution and digital filters, as for example obtained from the "Computer Vision/Image Processing lecture IP1". If no such course was taken, it is possible to obtain the required knowledge in self-study.

Literature:
Will be announced during the seminar.