English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Poster

The role of components in recognition across changes of view

MPS-Authors
/persons/resource/persons84062

Liter,  JC
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;
Max Planck Institute for Biological Cybernetics, Max Planck Society;

/persons/resource/persons83839

Bülthoff,  HH
Department Human Perception, Cognition and Action, Max Planck Institute for Biological Cybernetics, Max Planck Society;
Max Planck Institute for Biological Cybernetics, Max Planck Society;

External Resource
Fulltext (restricted access)
There are currently no full texts shared for your IP range.
Fulltext (public)
There are no public fulltexts stored in PuRe
Supplementary Material (public)
There is no public supplementary material available
Citation

Liter, J., & Bülthoff, H. (1996). The role of components in recognition across changes of view. Poster presented at 19th European Conference on Visual Perception (ECVP 1996), Strasbourg, France.


Cite as: https://hdl.handle.net/11858/00-001M-0000-0013-EB2E-9
Abstract
We examined the extent to which recognition of rotated objects is based on complex features, such as differently shaped components, when distractor objects may project images that are very similar to the studied images. The stimuli were unfamiliar computer-generated objects composed of four long thin components connected end to end. Each object was matched to a distractor object that had the same components arranged in a different order. The 3-D structure of the matched objects was otherwise the same---the lengths of the components in corresponding positions and the connection angles between them were identical. Subjects studied six target objects and then performed a 2AFC recognition task with each target and its matched distractor. Either the target, the distractor, or both were rotated ±40° about the vertical axis from the studied viewpoint. As expected, subjects chose the target object most often when it was not rotated (77.2 and 69.4 correct when the distractor was and was not rotated). However, subjects chose the target object 64.4 of the time when it was rotated and the distractor was not. In this case, the image of the distractor was more similar to the studied image than was the image of the target. Recognition of these objects was not based on pixel-based image similarity among studied and tested objects, but on a more general view-based similarity that takes into account a multitude of features, some more stable over changes of view than others. We discuss computational models that also rely on multiple view-based features.