Experiments in Texture Perception

Abstract

Visual textures may be described completely by their spatial frequency components. For one-dimensional textures whose luminance varies only along the X-axis of the display, the descriptive elements are gratings that have sinusoidal modulations of luminance. Although any arbitrary 1-dimensional 'blurred' texture may require a large number of sinusoidal components for its complete physical description, only 4 components are needed to create a texture that appears the same to the human observer. The human visual system seems does not act like a spectral analyser, but rather appears to process spatial frequency information by filtering operations, at least for 1-dimensional texture patterns. In the more general case, textures will have luminance distributions varying in both X and Y dimensions. A new graphics display is being built to test for the minimum number of spatial frequencies required to simulate 2-dimensional texture patterns. The apparatus will permit on-line control of the amplitude (contrast) of the (X,Y) frequency (Fourier) components that make up the texture displayed. The observer can generate a texture that appears identical to another having a different and more complex spatial frequency content. If it is found that only 4 spatial frequency components are necessary to simulate all 2-dimensional textures, one may design a scheme to transmit visual information about textures that offers considerable saving in channel capacity.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1975
Accession Number
ADA024975

Entities

People

  • Whitman Richards

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Biomedical
  • Space

DTIC Thesaurus Topics

  • Amplitude
  • Analyzers
  • Channel Capacity
  • Classification
  • Color Vision
  • Computers
  • Data Displays
  • Discrimination
  • Display Systems
  • Graphics
  • Modulation
  • Photographic Materials
  • Psychology
  • Security
  • Spectra
  • Two Dimensional
  • Visual Perception

Fields of Study

  • Physics

Readers

  • Computer Vision.
  • Radio communications and signal processing.