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

Learning viewpoint invariant object representations using a temporal coherence principle.

Wolfgang Einhäuser1, Jörg Hipp, Julian Eggert

  • 1Institute of Neuroinformatics, University & ETH Zürich, Zürich, Switzerland. wet@klab.caltech.edu

Biological Cybernetics
|July 16, 2005
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

Fixation duration on natural scenes is explained by memory encoding not processing demand.

Nature neuroscience·2026
Same author

Visual search for people wearing protective clothing in a forestial environment - Differences between gaze and behavior.

PloS one·2026
Same author

Mobile eye tracking in the real world: Best practices.

Journal of vision·2026
Same author

Task-irrelevant human and robot head movements bias gaze in humans who follow them through virtual reality.

Scientific reports·2026
Same author

Beyond the first glance: How human presence enhances visual entropy and promotes spatial learning.

PLoS computational biology·2026
Same author

Maternal antibiotic exposure-mediated alterations in basal, and allergen-induced lung function are associated with altered recruitment of eosinophils to the developing lung.

Frontiers in immunology·2026
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 demonstrates unsupervised learning for artificial vision systems. Unsupervised learning using temporal coherence principles enables viewpoint-invariant object recognition, even with distractors.

Area of Science:

  • Computer Vision
  • Computational Neuroscience
  • Machine Learning

Background:

  • Invariant object recognition is a key challenge for artificial vision systems, unlike the mammalian visual system.
  • Understanding biological visual systems reveals general coding principles that explain neuronal responses.
  • Transferring these biological principles to artificial systems can improve performance.

Purpose of the Study:

  • To investigate if unsupervised learning, based on biological coding principles, can achieve viewpoint-invariant object recognition in artificial systems.
  • To train artificial neural networks using a "stability" objective derived from temporal coherence.
  • To evaluate the system's ability to classify objects under varying viewpoints and with distractors.

Main Methods:

Related Experiment Videos

  • Training artificial neural network cells on unlabeled real-world images using a "stability" objective (a variant of temporal coherence).
  • Evaluating the resulting internal representations for viewpoint independence.
  • Testing classification performance on previously unseen viewpoints and in the presence of unlabeled distractor objects.

    • Main Results:

    • Trained cells formed representations largely independent of the viewpoint of the input stimulus.
    • This viewpoint-invariant representation generalized to novel viewpoints.
    • Classification performance improved compared to input patterns, even with distractor objects present.

    Conclusions:

    • Unsupervised learning, guided by general coding principles like temporal coherence, can facilitate robust object classification in artificial vision.
    • This approach enables recognition of objects undergoing complex transformations and without prior segmentation.
    • The findings suggest a pathway for developing more capable artificial visual systems inspired by biological principles.