Psychophysical measurement and modeling of human impressions
Prof. Dr. ir. Jean-Bernard Martens
Department of Industrial Design
Eindhoven University of Technology


There is a substantial economical interest in being able to measure (and possibly also predict) the effect of variations in technical parameters of image-based systems (such as TVs or VR-systems) on the quality of these systems, as perceived and experienced by end users. In this talk I concentrate on how quality and its underlying attributes can be measured with the help of human subjects. The theoretical models that underlie such measurements are discussed and it is demonstrated how the interactive multidimensional scaling (and regression) program XGms can be used to obtain insight into the relationship between these theoretical models and actual data. Most of the available data is taken from earlier work, where the effect of image processing and coding on image quality has been studied. I will also shortly discuss how this framework might be applied to more recent work on characterizing spatial interactions in virtual desktop environments (i.e., the new NWO-VIEW project "QUASID-Quantitative Spatial Interaction Design", conducted together with Robert van Liere from CWI).