Colour in Context
Research group
Computer Vision Center

Spatiochromatic information in natural scenes

Tom Troscianko, C. Alejandro Parraga, D.J. Tolhurst
Perception, European Conference on Visual Perception (ECVP´02), Volume 31, page 149b - 2002
IF: 1.314. area: PSYCHOLOGY.
Download the publication : p210.pdf [38Ko]  
To what extent does colour vision match the properties of natural scenes? Chromatic mechanisms are relatively more sensitive to low spatial frequencies than luminance mechanisms. Perhaps this reflects some general property of natural scenes, but empirical evidence argues against this. However, since there is separate evidence that colour vision evolved for frugivory, we asked whether the properties of scenes of fruit and foliage differ from those of general scenes. The results suggested that such scenes are indeed optimally signalled by primate spatiochromatic vision, but only for the red - green system (Párraga et al, 2002 Current Biology 12 483 - 487). What about the yellow - blue system? Its spatial properties are similar to that of the red - green system. However, we have found no spatial optimisation for foliage encoding. The problem with this system is that shadows are bluer than highlights, and are therefore not removed from dappled scenes. The only situation in which there is evidence of spatial optimality is when patches of blue sky are visible through the leaf canopy. Thus, spatial encoding by the yellow - blue system may be optimised for distinguishing between foliage and sky, or perhaps for an entirely different, yet undiscovered, class of tasks.

Images and movies

 

BibTex references

@InProceedings\{TPT2002,
  author       = "Tom Troscianko and C. Alejandro Parraga and D.J. Tolhurst",
  title        = "Spatiochromatic information in natural scenes",
  booktitle    = "Perception, European Conference on Visual Perception (ECVP\´02)",
  volume       = "31",
  pages        = "149b",
  year         = "2002",
  abstract     = "To what extent does colour vision match the properties of natural scenes? Chromatic mechanisms are relatively more sensitive to low spatial frequencies than luminance mechanisms. Perhaps this reflects some general property of natural scenes, but empirical evidence argues against this. However, since there is separate evidence that colour vision evolved for frugivory, we asked whether the properties of scenes of fruit and foliage differ from those of general scenes. The results suggested that such scenes are indeed optimally signalled by primate spatiochromatic vision, but only for the red - green system (P\'arraga et al, 2002 Current Biology 12 483 - 487). What about the yellow - blue system? Its spatial properties are similar to that of the red - green system. However, we have found no spatial optimisation for foliage encoding. The problem with this system is that shadows are bluer than highlights, and are therefore not removed from dappled scenes. The only situation in which there is evidence of spatial optimality is when patches of blue sky are visible through the leaf canopy. Thus, spatial encoding by the yellow - blue system may be optimised for distinguishing between foliage and sky, or perhaps for an entirely different, yet undiscovered, class of tasks.",
  ifactor      = "1.314",
  area         = "PSYCHOLOGY",
  url          = "http://cat.cvc.uab.es/Public/Publications/2002/TPT2002"
}

Other publications in the database

 © 2008 Colour in context Group | Computer Vision Center. All rights reserved | Contact webmaster |  Last updated: Monday 11 May 2009     eXTReMe Tracker