POPPerception On Purpose

Eye vergence plays a crucial role in stereopsis. We addressed computational issues in accordance with the biological mechanism of eye vergence and its role in estimating depth from binocular cues.

Contributing partners : INRIA, University of Coimbra

An algorithm that provides disparities in accordance with the surface properties of the scene under consideration has been proposed.

Contributing partner : INRIA.

We addressed detection and localizion of objects that are both seen and heard. We capitalized on a human-like configuration of sensors (binaural and binocular). We proposed a probabilistic generative model that captures the relations between audio and visual observations.

Contributing partners : INRIA, University of Sheffield.

We have explored the new idea of auditory gist, i.e., the perception of very brief segments of speech.

Contributing partners : University of Sheffield and University of Osnabrueck.

POP has constructed novel algorithms for the robust localisation of sound sources in multisource environments.

Contributing partner : University of Sheffield

POP has considered the new possibilities that arise when moving beyond the static perceiver assumptions of previous hearing research.

Contributing partner : University of Sheffield

We developed several prototypes of an audi-visual robot head with four degrees of freedom. It allows eye-gaze, eye-vergence, as well as rapid neck rotations. It can hold hold a dummy head with special accoustic properties.

Contributing partner : University of Coimbra.

EEG and MEG are severely affected by electric and magnetic artefacts induced by eye movements. In POP we developed, validated and applied a method to remove eye-movement artefacts from EEG and MEG recordings.

Contributing partners : University Hospital Hamburg-Eppendorf, University of Osnabrueck.

In the POP project, we have been able to provide evidence in several EEG studies in humans that coherence of neural signals may be an essential mechanism supporting multisensory perception.

Contributing partners : University Hospital Hamburg-Eppendorf.

JAMF is a flexible modelling environment which allows non-experts to develop models of sensory processing and attention.

Contributing partners : University of Osnabrück, University of Coimbra, University of Sheffield.