Phantom surface captures stereopsis

1
Phantom surface captures stereopsis

• Jukka Hakkinen , Gote Nyman
• Vision Research 41 (2001) 187199

Wijb Sommer Skadi van der Meer Tijmen van Wettum
2
Introduction

• We construct 3D images with diffrences from views
  of the left and right eye.
• Non-corresponging areas are more than meaningless
  noice.
• We can interpred them as "half occuding Images.

3
Old exeriments suggested that half-occlusion is
independant from conventional stereotopsis.
Thus interaction of binocular and monocular
elements is an intresting reasearch topic.
4

• A periodic textured wallpaper can function as a
  sensitive indicator of interactions in
  stereotopises.
• If the wallpaper is made of lots of small dots
  the 3D position of those dots can be changed
  easily.
• The exeriment was to find out phantom surface
  could capture the wallpaper dots

5
(No Transcript)
6

• They also used phantom Kanzizsa surfaces.
• If a three-dimensional illusory surface is
  perceived in a phantom Kanizsa figure it is
  either due to a half-occlusion interpretation
  made possible by the monocular areas or because
  the corner areas are matchable features.

7
The Experiment

• The subjects had to tell if the dots in the
  middle appeared nearer or the bar under the
  picture.
• Each picture had a crossed and an un-crossed
  version

8
Results

• Depth assymetric
• Near depth all the near central are pulled in
  front of the background.
• Far depth only the dots inside the corners are
  located futher away
• The cross increased the clarity of the illusion

9
Results

• The near capture seems particularly strong with
  phantom Kanizsa surfaces.
• In all first and last phantom surface (full
  squarelines) configuration those with the
  "cross" are sronger preceived. But the 2nd
  (removed 1 vertical) it effect diminished.

10
General Discussion