Computational Visual Perception



Time and place


How do humans perceive the visual world? How can we build computational models to mimic this human perception? And how can we validate those computational models? This course is designed as an introduction to enable you to build computational models for human visual perception. It will therefore provide an introduction into the human visual system building on the course on cognitive neuroscience for AI developers. You will learn how the human eye and brain process visual input and what we currently know about the ventral visual stream. We will look at computational models for all different levels of visual processing and discuss how well they measure behavioral data. This lecture is designed to be at the intersection of Computer Science (Computer Vision and Graphics) and Cognitive Neuroscience. After an initial introductory phase, you will in small teams (2-3 students) perform smallprojects to build prototypes for computational models for visual processing, reproduce recent scientific results or experiment with existing models. In parallel with the project phase we will read and discuss four recent research papers studying potential computational models and investigate how we can evaluate computational models.

Additional information

Expected participants: 20