Contents

In this seminar, we study the problem of visual object recogntion. While humans can effortlessly recognize thousands of objects despite changes in illumination, pose, articulation, etc., it has so far not been possible to mimic similar competence in a computer. What are the reasons for this? What makes the problem difficult? To what extent do we understand nature's solution or the way our brains arrive at it through learning during our development? What are promising new approaches for building object recognition systems for computers? What are the theoretical and practical challenges? To answer these questions, we will read important and recent papers from the fields of computer vision, visual psychophysics, and visual neuroscience, which participants will present to each other. Other students and the instructors will give feedback on the presentations.

Requirements

Open-minded students from a number of disciplines including Physics, Computer Science, Biology, and Psychology are welcome. Due to the interdisciplinary nature of the course no specific prerequisite knowledge is required.