Visual Perception Engineering
|Dozent: ||Dr.-Ing. Michael Dorr|
|Assistent:||Mikhail Startsev, Ioannis Agtzidis|
|Zielgruppe:||Wahlmodul zur fachlichen Vertiefung, MSEI, MSNE|
|Umfang:||2 VO / 2 UE|
|Prüfung:||schriftlich, 60 Min.|
|Zeit & Ort:|
Fundamentals of the human visual system: anatomy and functional description; Introduction to visual psychophysics: signal detection theory and adaptive Bayesian testing; Laboratory setup; Display technology and calibration: luminance, contrast, and temporal characteristics; Eye-tracking methodology and data analysis; Gaze-based interfaces: gaze as explicit and implicit input modality; Multi-resolutional gaze-contingent displays to study visual perception; Perceptual rendering and image quality metrics; Natural image statistics; Models of saliency and attention; Perception-inspired computer vision algorithms
Tutorial: psychophysical testing methods, evaluation of models of eye movements and attention, perceptual image quality metrics
After completion of this module, students know core principles of information processing in the human visual system, and they understand how visual signals can be efficiently encoded and transmitted exploiting these principles.
Students know how visual performance can be measured in the laboratory, and are able to design and conduct basic psychophysical experiments.
They have acquired knowledge of different techniques to track attention of an observer, and are able to build systems that react to the observer's attentional state.
They can implement simulations and models of visual processing and attention for naturalistic images and videos. They are able to analyse state-of-the-art perception-inspired algorithms and systems such as pipelines for action recognition in videos.
The following literature is recommended:
- Lu, Dosher, Visual Psychophysics, The MIT Press 2014.
- Holmqvist et al., Eye tracking: A comprehensive guide to methods and measures, Oxford University Press 2010.