392167 Automatic Human Behaviour Analysis Using Machine Learning and Psychological Methods (Pj) (SoSe 2020)

In this project, students will develop computational models to detect human behaviours (e.g. engagement, interest, boredom, etc.) during human-robot or human-computer interaction. To develop these models, the students will combine psychological knowledge about human behaviours with computational methods for behavioural signal processing. The models will be developed on available datasets either by using raw multimodal sensor data or by using annotated behavioural cues such as facial expressions and body pose. The developed models will be later tested using new data collected through self-designed user studies in human-robot or human-computer interaction. The user studies will be designed based on psychological principles and practices.

NOTE: Only a maximum of 10 participants is possible. A first-come-first-serve procedure will be followed. The first 10 students who attend the first session(s) will be able to participate. If more than 10 students participate in the first session(s), then a selection process will be conducted, in which the students will be asked to prepare and submit project proposals in groups. The three best project proposals will be selected, and these project groups will be able to continue their participation in future sessions (exceptions may be possible on a case-by-case basis).

All sessions will take place live via videoconferencing using the software Zoom. The links to the Zoom videoconference will be emailed to participants at least one day in advance. The first session will take place on 22.04.2020 at 10:15 am.