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

Visual recognition and inference using dynamic overcomplete sparse learning.

Joseph F Murray1, Kenneth Kreutz-Delgado

  • 1Massachusetts Institute of Technology, Brain and Cognitive Sciences Department, Cambridge, MA 02139, USA. murrayjf@mit.edu

Neural Computation
|July 26, 2007
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

Bridging M/EEG Source Imaging and Independent Component Analysis Frameworks Using Biologically Inspired Sparsity Priors.

Neural computation·2021
Same author

What Can Local Transfer Entropy Tell Us about Phase-Amplitude Coupling in Electrophysiological Signals?

Entropy (Basel, Switzerland)·2020
Same author

The open EEGLAB portal Interface: High-Performance computing with EEGLAB.

NeuroImage·2020
Same author

Unraveling the spatiotemporal brain dynamics during a simulated reach-to-eat task.

NeuroImage·2018
Same author

Fast and robust Block-Sparse Bayesian learning for EEG source imaging.

NeuroImage·2018
Same author

Validation of periodic fMRI signals in response to wearable tactile stimulation.

NeuroImage·2017
Same journal

A Model-Free Reinforcement Learning Implementation of Decision Making Under Uncertainty by Sequential Sampling.

Neural computation·2026
Same journal

DROP: Distributional and Regular Optimism and Pessimism for Reinforcement Learning.

Neural computation·2026
Same journal

Hierarchical Active Inference Using Successor Representations.

Neural computation·2026
Same journal

W-Kernel and Its Principal Space for Frequentist Evaluation of Bayesian Estimators.

Neural computation·2026
Same journal

A Hidden Markov Model-Inspired Sequence Classification Method for Hyperdimensional Computing.

Neural computation·2026
Same journal

Sparse Graphical Modeling for Electrophysiological Phase-Based Connectivity Using Circular Statistics.

Neural computation·2026
See all related articles

This study introduces a hierarchical model for visual recognition and inference tasks, inspired by biological vision. The approach enhances image recognition, especially in cluttered or occluded scenes, by using sparse coding and overcomplete representations.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Computational Neuroscience

Background:

  • Biological vision principles guide the development of computational models.
  • Sparse coding and overcomplete representations are increasingly recognized for their utility in visual tasks.

Purpose of the Study:

  • To develop a hierarchical architecture and learning algorithm for visual recognition and inference.
  • To create a simplified world model (SWM) using variational approximation for sparse coding.
  • To improve performance on complex visual tasks like object recognition with occlusion.

Main Methods:

  • A stochastic generative world model is posited, simplified via variational approximation for sparse coding.
  • An overcomplete dictionary learning algorithm is used for sparse image coding.

Related Experiment Videos

  • A dynamic multilayer network with feedforward, feedback, and lateral connections is trained to approximate the SWM.
  • A variant of back-propagation-through-time is employed for learning, optimizing for sparse layer updates.

    • Main Results:

    • Experiments with rotated objects demonstrate capabilities in visual imagination, occlusion reconstruction, and segmentation.
    • Increasing the degree of overcompleteness in the dictionary learning stage improved recognition accuracy in cluttered and occluded scenes.
    • The dynamic network successfully approximates the simplified world model.

    Conclusions:

    • The proposed hierarchical architecture and learning algorithm are effective for various visual inference tasks.
    • Sparse coding and overcomplete representations are crucial for robust visual recognition, particularly in challenging conditions.
    • The model offers a biologically plausible approach to artificial visual systems.