Principal Components of Natural Images: An Analytical Solution

Abstract

The structure of receptive fields in the visual cortex is believed to be shaped by unsupervised learning. A simple variant of unsupervised learning is the extraction of principal components. In this paper, we derived analytically the form of the principal components of natural images. This derivation relies on results about the covariance matrix of natural images. Our results predict both the shapes and the phases of the receptive fields. We also compared our results to numerical simulation results. Finally the biological relevance of our results is discussed.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 17, 1993
Accession Number
ADA264800

Entities

People

  • Harel Shouval
  • Yong Liu

Organizations

  • Brown University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Brain
  • Covariance
  • Eigenvalues
  • Equations
  • Identification
  • Information Processing
  • Information Science
  • Neural Networks
  • Pattern Recognition
  • Preprocessing
  • Recognition
  • Security
  • Self Organizing Systems
  • Simulations
  • Two Dimensional
  • Unsupervised Machine Learning
  • Visual Cortex

Readers

  • Approximation Theory.
  • Finite Element Method (FEM) for solving Partial Differential Equations (PDEs)
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms