Learning about human edge perception through noise masking experiments

Project Description

It is widely known that edges (e.g. object boundaries) serve as a major source of information to the human visual system. We define an edge as a luminance discontinuity in an image which is localized in space and extends along one axis. To test human edge perception experimentally, different tasks have been proposed such as edge localization [1] and edge polarity judgments in noise [2]. However, it is still debated to which extent human edge perception relies on processing in a wide range of so-called spatial-frequency-selective channels. To better answer this question, we want to probe human edge perception in noise masking experiments in which we specifically interfere with the presumed underlying mechanisms. The aim of this project will be to implement and conduct a psychophysical experiment in our laboratory, and to analyze and interpret the resulting data with statistical methods.

Students will have to

• scan and read papers from the edge perception literature
• write code to conduct and evaluate a psychophysical experiment

Students will learn

• about fundamental mechanisms of visual processing
• how to implement and evaluate psychophysical experiments with statistical methods

Requirements

• Good programming skills in Python
• Good English proficiency

If you are interested, please contact Lynn Schmittwilken (L.Schmittwilken@tu-berlin.de)