Date: Wednesday 26 May 2021
Venue: Via Zoom
Speaker: Ms Tram Nguyen
Our natural world is often populated with clusters of visual features and objects that share similar characteristics (e.g., leaves on a tree, a flock of birds, a group of people). Ensemble perception theories ask whether we exploit such cluster similarity to overcome known attentional limits in processing multiple objects.
Specifically, researchers have suggested that our visual system operates in a statistical fashion that can quickly extract the average, as well as dispersion, of redundant visual characteristics. Coding visual clusters as ensembles together with their summary statistics is thought to be effective in terms of computational cost, allowing rapid visual analysis of scene ‘gist’, for example.
While ensemble perception of low-level features (e.g., colour and orientation) has been well established, ensemble perception of high-level objects (e.g., human faces and bodies) remains an active area of research. Several questions related to high-level ensemble coding remain unanswered.
Do we need to effortfully calculate average information or does such ‘averageness’ emerge automatically? Do summary statistics take into account all of the presented information or only part of it? Does unattended information contribute to the averaging process? How can we reconcile the limits of focused attention with evidence of high-level ensemble processing? In this presentation, recent data on high-level ensemble perception will be discussed using a task to estimate the average walking speed of a human crowd.
'Ensemble Perception of Human Crowds' is a research webinar to be hosted by the Department of Cognitive Science within the Faculty of Media and Knowledge Sciences within the University of Malta.
Kindly confirm your participation by sending an email.