Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Inferring perceptual saliency fields from viewpoint-dependent recognition data.

F Cutzu1, M Tarr

  • 1Department of Computer Science, Sanford Fleming Bldg, University of Toronto, Toronto, Ontario M5S 3G4, Canada. florin@vis.toronto.edu

Neural Computation
|July 29, 1999
PubMed
Summary

We developed a new algorithm to identify important features on 3D objects by analyzing how people perceive them. This method helps understand object perception by finding salient regions on surfaces.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Calcium as a modulator of photosensitized killing of H9c2 cardiac cells.

Photochemistry and photobiology·2001
Same author

Photosensitization-induced calcium overload in cardiac cells: direct link to membrane permeabilization and calcium influx.

Photochemistry and photobiology·2001
Same author

[Alteration of nitric oxide production during arterial myocardial revascularization using extracorporeal circulation. Quantitative measurement of EDNO production by the internal mammary artery bypass graft].

Orvosi hetilap·2001
Same author

Representation of object similarity in human vision: psychophysics and a computational model.

Vision research·1998
Same author

GH3 cells, ionic currents and cell killing: photomodification sensitized by Rose Bengal.

Photochemistry and photobiology·1998
Same author

Photomodification of cardiac membrane: chaotic currents and high conductance states in isolated patches.

Photochemistry and photobiology·1998

Area of Science:

  • Computer Vision
  • Human Perception
  • Computational Geometry

Background:

  • Understanding human 3D object perception is challenging.
  • Existing theories often rely on arbitrary features.
  • Predicting object view quality or similarity from features is an inverse problem.

Purpose of the Study:

  • To compute relative perceptual saliencies of 3D object features.
  • To address the inverse problem of predicting view quality/similarity.
  • To identify object structures crucial for human perception.

Main Methods:

  • Developed a regularization method based on a linear model.
  • Assumed perceptual salience varies slowly across object surfaces.
  • Utilized goodness-of-view scores and perceptual similarities from multiple viewpoints.

Related Experiment Videos

Main Results:

  • Successfully computed perceptual saliencies for 3D object features.
  • The algorithm identifies salient regions on object surfaces.
  • Empirically validated the importance of derived salient regions.

Conclusions:

  • The proposed algorithm provides a principled approach to understanding 3D object perception.
  • Identified salient regions offer insights into human visual processing of objects.
  • This method moves beyond ad hoc features in perceptual theories.