Situated computer vision [Elektronische Ressource] / Sven Wachsmuth

Situated Computer VisionSven WachsmuthHabilitationsschrift(Tag der Habilitation: 30.1.2009)Universit at BielefeldTechnische Fakult atMarch 24, 20102Contents1 Situated Perception 111.1 Perspectives on computer vision . . . . . . . . . . . . . . . . . 121.2 Situation models . . . . . . . . . . . . . . . . . . . . . . . . . 151.2.1 Storage and retrieval structures . . . . . . . . . . . . . 171.2.2 Situation models as a dynamical representation . . . . 191.3 Context in Human vision . . . . . . . . . . . . . . . . . . . . . 211.3.1 Results from eye-tracking experiments . . . . . . . . . 221.3.2 The object-detection paradigm . . . . . . . . . . . . . 231.3.3 Neurophysiological results . . . . . . . . . . . . . . . . 241.4 Summary and conclusion . . . . . . . . . . . . . . . . . . . . . 272 Perception of Scenes 292.1 Why context? . . . . . . . . . . . . . . . . . . . . . . . . . . . 302.1.1 Aspects of contextual modeling . . . . . . . . . . . . . 332.1.2 Contextual modeling for scene understanding . . . . . 352.1.3 Con modeling for system control . . . . . . . . . 382.2 Recognizing global scene contexts . . . . . . . . . . . . . . . . 412.2.1 Holistic scene classi cation . . . . . . . . . . . . . . . . 432.2.2 Scenes as a con guration of parts . . . . . . . . . . . . 522.3 Using context in object recognition . . . . . . . . . . . . . . . 652.3.1 Combining holistic context and object detection . . . . 672.3.