Spatial structure and the perceived motion of ob jects of different colors

Spatial structure and the perceived motion of ob jects of different colors 92 Spatial structure and the perceived motion of ob jects of different colors BART FARELL Institute for Sensory Research, Syracuse University, Syracuse, NY 13244-5290, USA The nature of color's contribution to the perception of motion is not completely clear: are motion signals from chromatic and luminance stimuli processed independently or not? Plaids containing 3 component gratings permit study of interactions between ID luminance and chromatic stimuli with identical velocity vectors. These plaids show that whether processing of luminance and chromatic components is independent or interactive depends entirely on spatial variables. Observers see a coherent multicolored plaid only when luminance components have a higher spatial frequency than chromatic components and components of each type have a common orientation. Otherwise luminance and chromatic components do not cohere. The role of luminance edges and relative phase was investigated by combining a chromatic sinusoid and a luminance plaid composed of square waves minus the fundamental. At 0° relative phase, the 3 components are seen to cohere, as found before. Remarkably, a 90° shift in relative phase made the chromatic content of the pattern invisible, leaving behind a thoroughly achromatic plaid. The dependence of perceived motion on spatial parameters suggests that luminance-chromatic interactions are an attempt by the visual system to exploit the expected correlation between luminance and chromatic spatial distributions across objects, while also heavily weighting the contribution of moving luminance edges. Texture processing and image segmentation in man and machines: a unified theory TERRY CAELLI Department of Computer Science, The University of Melbourne, Parkville. Victoria, Australia 3052 Over the past thirty years of investigating the processes that underlie human and machine texture processing and image segmentation a number of computational processes have been shown to be sufficient to attain region identification which confonns to what humans perceive. We attempt to identify, and enumerate the characteristics of these processes and to show how they are common to early processing in both human and machine vision. Finally, we contrast texture or region classification with segmentation problems and argue that the latter is much more fundamental to early visual processing - involving, for stable solutions, more constraints than previously thought in both literatures. Put simply, the main processes involved in texture classification or region classification or segmentation are: ( 1 ) an adaptive form of decomposition - usually expressed either in the form of local auto-correlators, adaptive filter (feature) profiles or the differential weighting of fixed filters; (2) a measure of filter output to 'evidence' feature presence - in particular, invariant region and boundary predicates; (3) a form of constrained clustering with at least one constraint due to spatial contiguity. Evidence for these processes will be given from a variety of biological and machine vision investigations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Spatial Vision (continued as Seeing & Perceiving from 2010) Brill

Spatial structure and the perceived motion of ob jects of different colors

Free
1 page

Loading next page...
1 Page
 
/lp/brill/spatial-structure-and-the-perceived-motion-of-ob-jects-of-different-A3P7EFvj95
Publisher
BRILL
Copyright
© 1993 Koninklijke Brill NV, Leiden, The Netherlands
ISSN
0169-1015
eISSN
1568-5683
D.O.I.
10.1163/156856893X00270
Publisher site
See Article on Publisher Site

Abstract

92 Spatial structure and the perceived motion of ob jects of different colors BART FARELL Institute for Sensory Research, Syracuse University, Syracuse, NY 13244-5290, USA The nature of color's contribution to the perception of motion is not completely clear: are motion signals from chromatic and luminance stimuli processed independently or not? Plaids containing 3 component gratings permit study of interactions between ID luminance and chromatic stimuli with identical velocity vectors. These plaids show that whether processing of luminance and chromatic components is independent or interactive depends entirely on spatial variables. Observers see a coherent multicolored plaid only when luminance components have a higher spatial frequency than chromatic components and components of each type have a common orientation. Otherwise luminance and chromatic components do not cohere. The role of luminance edges and relative phase was investigated by combining a chromatic sinusoid and a luminance plaid composed of square waves minus the fundamental. At 0° relative phase, the 3 components are seen to cohere, as found before. Remarkably, a 90° shift in relative phase made the chromatic content of the pattern invisible, leaving behind a thoroughly achromatic plaid. The dependence of perceived motion on spatial parameters suggests that luminance-chromatic interactions are an attempt by the visual system to exploit the expected correlation between luminance and chromatic spatial distributions across objects, while also heavily weighting the contribution of moving luminance edges. Texture processing and image segmentation in man and machines: a unified theory TERRY CAELLI Department of Computer Science, The University of Melbourne, Parkville. Victoria, Australia 3052 Over the past thirty years of investigating the processes that underlie human and machine texture processing and image segmentation a number of computational processes have been shown to be sufficient to attain region identification which confonns to what humans perceive. We attempt to identify, and enumerate the characteristics of these processes and to show how they are common to early processing in both human and machine vision. Finally, we contrast texture or region classification with segmentation problems and argue that the latter is much more fundamental to early visual processing - involving, for stable solutions, more constraints than previously thought in both literatures. Put simply, the main processes involved in texture classification or region classification or segmentation are: ( 1 ) an adaptive form of decomposition - usually expressed either in the form of local auto-correlators, adaptive filter (feature) profiles or the differential weighting of fixed filters; (2) a measure of filter output to 'evidence' feature presence - in particular, invariant region and boundary predicates; (3) a form of constrained clustering with at least one constraint due to spatial contiguity. Evidence for these processes will be given from a variety of biological and machine vision investigations.

Journal

Spatial Vision (continued as Seeing & Perceiving from 2010)Brill

Published: Jan 1, 1993

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off