Local contour integration depends on alignment and orientation of global shapes
-
1
University of Bremen, Department for Human Neurobiology and Bernstein Group for Computational Neuroscience, Germany
-
2
University of Bremen, Department for Neurophysics and Bernstein Group for Computational Neuroscience, Institute for Theoretical Physics, Germany
Shape perception in natural scenes relies on integration processes which construct a representation from local image features. An example is contour integration, grouping nearly collinear, aligned edge elements into coherent forms. At the same time, perception is strongly modulated by the global context of a shape: in a previous experiment, we found that reaction times for contours forming horizontal ellipses were shorter than for vertical or oblique ellipses. This horizontal bias in contour perception was also reflected by differences in the EEG potentials recorded from parietal and occipital electrodes.
In this contribution, we first quantify which geometrical factors of the global context contribute to the horizontal bias. Second, we investigate whether the horizontal bias is an effect common to feature integration, or specific to contour integration of oriented edge elements.
In psychophysical experiments observers had to detect contours of locally aligned edge elements, embedded in a background of randomly oriented distractors. The global form of the contours was either a horizontal or a vertical ellipse. To assess differences in the contour percepts, a staircase procedure was used to determine the detection thresholds for horizontal and vertical ellipses with respect to a control parameter.
In our first experiment, the control parameter was the relative alignment (jitter) of the edges to the contour path. We independently varied three global, contextual factors: (A) eccentricity of the center of the contour from the fixation spot (6.68 arcdeg vs. 3.34 arcdeg), (B) radial versus tangential alignment relative to the fixation spot, and (C) display aspect ratio (4:3, horizontal versus 3:4, vertical). In addition to the horizontal bias, only factor (B) had a significant influence on the thresholds; ellipses aligned radially to the fixation spot were easier to detect than tangentially aligned ellipses. The relative strength of this effect depends on the eccentricity of the contours. In an additional task reaction times were measured. Consistently, responses for horizontal stimuli were significantly faster than for vertical stimuli.
In our second experiment, the orientations of all edge elements were random. Contours were now defined by a set of edges with higher contrast, which was the control parameter for the staircase that determined the visibility of their elliptic shapes. We found no perception biases and no reaction time differences for both, horizontal or radially aligned contours.
In summary, we found that shape perception is modulated by both, the alignment of contours to the fixation spot, and their global shape (horizontal versus vertical). Surprisingly, these effects specifically target integration processes of local oriented edge elements.
Acknowledgements
Supported by BMBF / National Bernstein Network for Computational Neuroscience (Bernstein Group Bremen, Förderkennzeichen 01GQ0705)
Keywords:
computational neuroscience
Conference:
Bernstein Conference on Computational Neuroscience, Berlin, Germany, 27 Sep - 1 Oct, 2010.
Presentation Type:
Presentation
Topic:
Bernstein Conference on Computational Neuroscience
Citation:
Schipper
M,
Teiwes
V,
Fahle
M and
Ernst
U
(2010). Local contour integration depends on alignment and orientation of global shapes.
Front. Comput. Neurosci.
Conference Abstract:
Bernstein Conference on Computational Neuroscience.
doi: 10.3389/conf.fncom.2010.51.00108
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
08 Sep 2010;
Published Online:
23 Sep 2010.
*
Correspondence:
Dr. Marc Schipper, University of Bremen, Department for Human Neurobiology and Bernstein Group for Computational Neuroscience, Bremen, Germany, marc.schipper@uni-bremen.de