Predicting human behavior in smart environments : theory and application to gaze prediction

Redwan Abdo Abdullah Mohammed

doi:10.18453/rosdok_id00001865

Redwan Abdo Abdullah Mohammed

Predicting human behavior in smart environments : theory and application to gaze prediction

Universität Rostock, 2017

https://doi.org/10.18453/rosdok_id00001865

Abstract: Predicting human behavior is desirable in many application scenarios in smart environments. The existing models for eye movements do not take contextual factors into account. This addressed in this thesis using a systematic machine-learning approach, where user profiles for eye movements behaviors are learned from data. In addition, a theoretical innovation is presented, which goes beyond pure data analysis. The thesis proposed the modeling of eye movements as a Markov Decision Processes. It uses Inverse Reinforcement Learning paradigm to infer the user eye movements behaviors.

Dissertation Freier Zugang

Titel:

Predicting human behavior in smart environments: theory and application to gaze prediction

Beteiligte Personen:

Redwan Abdo Abdullah Mohammed[VerfasserIn]
	112675353X
Oliver Staadt , Prof. Dr. Sc. techn.[AkademischeR BetreuerIn]
	123136539
	University of Rostock, Visual Computing Group, Institute of Computer Science
Thomas Kirste , Prof. Dr.-Ing.[AkademischeR BetreuerIn]
	137244754
	University of Rostock, Mobile Multimedia Information Systems Group, Institute of Computer Science
Henning Sprekeler , Prof. Dr.[AkademischeR BetreuerIn]
	138094357
	Technical University Berlin, Modelling of Cognitive Processes Group, Institute of Software Engineering and Theoretical Computer Science
Lars Schwabe , Dr. rer. nat.[AkademischeR BetreuerIn]
	141827793
	Data Insight Lab, Lufthansa Industry Solutions

Beteiligte Körperschaften:

Universität Rostock, Fakultät für Informatik und Elektrotechnik[Grad-verleihende Institution]
	10085032-7

Zusammenfassung:

Predicting human behavior is desirable in many application scenarios in smart environments. The existing models for eye movements do not take contextual factors into account. This addressed in this thesis using a systematic machine-learning approach, where user profiles for eye movements behaviors are learned from data. In addition, a theoretical innovation is presented, which goes beyond pure data analysis. The thesis proposed the modeling of eye movements as a Markov Decision Processes. It uses Inverse Reinforcement Learning paradigm to infer the user eye movements behaviors. [Englisch]

Dokumenttyp:

Dissertation

Einrichtung:

Fakultät für Informatik und Elektrotechnik

Sprache:

Englisch

Sachgruppe der DNB:

004 Informatik

Veröffentlichung /
Entstehung:

Rostock

Rostock: Universität Rostock

2017