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

A self-organizing multiple-view representation of 3D objects.

S Edelman1, D Weinshall

  • 1Center for Biological Information Processing, Department of Brain and Cognitive Sciences, MIT, Cambridge, MA 02139.

Biological Cybernetics
|January 1, 1991
PubMed
Summary
This summary is machine-generated.

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

Challenges faced by physicians when discussing the Type 2 diabetes diagnosis with patients: insights from a cross-national study (IntroDia<sup>®</sup> ).

Diabetic medicine : a journal of the British Diabetic Association·2017
Same author

Oxytocin receptor and vasopressin receptor 1a genes are respectively associated with emotional and cognitive empathy.

Hormones and behavior·2014
Same author

Control of type 2 diabetes after 1 year of laparoscopic adjustable gastric banding in the helping evaluate reduction in obesity (HERO) study.

Diabetes, obesity & metabolism·2014
Same author

Frailty and osteoporosis in older women--a prospective study.

Osteoporosis international : a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA·2013
Same author

Competitive learning in biological and artificial neural computation.

Trends in cognitive sciences·2011
Same author

Computational theories of object recognition.

Trends in cognitive sciences·2011
Same journal

Harmonic memory in phasor neural networks.

Biological cybernetics·2026
Same journal

Correction: Decreased spinal inhibition leads to undiversified locomotor patterns.

Biological cybernetics·2026
Same journal

Foundational issues of network models in biology.

Biological cybernetics·2026
Same journal

Dynamical mechanisms for coordinating long-term working memory based on the precision of spike-timing in cortical neurons.

Biological cybernetics·2026
Same journal

Distinct dopaminergic spike-timing-dependent plasticity rules are suited to different functional roles.

Biological cybernetics·2026
Same journal

Fluctuation-response relations for a two-stage population of spiking neurons stimulated by common noise.

Biological cybernetics·2026
See all related articles

This study shows a two-layer neural network can learn compact 3D object representations from multiple 2D views. The network generalized well to new viewpoints, mimicking human perception.

Area of Science:

  • Artificial Intelligence
  • Computer Vision
  • Cognitive Science

Background:

  • Representing 3D objects from 2D views is crucial for AI and understanding human perception.
  • Existing methods often require extensive labeled data or struggle with generalization.

Purpose of the Study:

  • To investigate a novel neural network architecture for 3D object representation using multiple 2D views.
  • To assess the network's ability to learn compact representations and generalize to unseen viewpoints.
  • To compare the network's performance with human psychophysical data.

Main Methods:

  • Utilized a two-layer network of thresholded summation units.
  • Employed unsupervised Hebbian relaxation for network training.
  • Trained the network on ten distinct objects from various viewpoints.

Related Experiment Videos

  • Evaluated generalization on novel views and simulated psychophysical experiments.

    • Main Results:

    • The network successfully learned to recognize objects from different viewpoints.
    • Compact representations of input views emerged during training.
    • Demonstrated substantial generalization capability on novel views.
    • Network behavior qualitatively matched human subject performance in psychophysical simulations.

    Conclusions:

    • A two-layer network effectively represents 3D objects using multiple 2D views.
    • Unsupervised learning enables compact representations and strong generalization.
    • The model provides insights into potential mechanisms of human 3D object recognition.