Probing and modelling object-centric dynamic scene processing in the human brain (B03 (ehem. 09))

• Subject Area: Cognitive, Systems and Behavioural Neurobiology ; Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing ; Molecular and Cellular Neurology and Neuropathology
• Term: since 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 276693517

Project B3 will investigate how the Gestalt principle of motion-based segmentation and invariant neural representations can support robust object-centric visual perception of dynamic, natural scenes. We will develop neural network models for motion-based object segmentation that align with biological vision constraints, and verify how well they match human vision both behaviorally through psychophysical experiments and neurally based on fMRI recordings and neural measurements from B1 and B2. Our goal is to identify potential solutions regarding the algorithms and representations in the visual brain that could facilitate stable object perception and few-shot generalisation despite changes in appearance variables like lighting conditions or viewing angles.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1233: Robust Vision - Inference Principles and Neural Mechanisms
• Applicant Institution: Eberhard Karls Universität Tübingen
• Project Heads: Professor Andreas Bartels, Ph.D.; Professor Dr. Matthias Bethge; Professor Michael Black, Ph.D., until 12/2024