Examples of images with rich monocular features and structures of various kinds.
This is part of a web site presenting some ideas about limitations in current
AI/Robotic vision systems, and possibly also inadequate current theories about neural
mechanisms required for vision.
This page is best viewed when linked from this page:
"How to look at what you see."
How many of these can be detected by AI systems and used in interpreting
images, e.g. aiding fusion in stereo pairs without having to fuse at the
pixel level as in Julesz random dot stereograms?
Here we use them only for "monocular" tasks.
Aaron Sloman
http://www.cs.bham.ac.uk/~axs
__________________________________________________________________________________________
Small extracts from the larger images that follow.
For each of the next eight images see if you can find which of the images below it
came from, and where it was in the original.
Hint: The source of one of the smaller images is particularly difficult to
locate because of the change of scale.
(The difficulty of the task may depend on the resolution and quality of your screen.)
__________________________________________________________________________________________
-
-
-
__________________________________________________________________________________________
If you wish to see the originals below in a separate window
Click Here
The originals -- and some additional images
For each of these, using Right-click and "View Image" you should be able to view the
original image expanded, which may be too large to fit on your display without some
compression.
__________________________________________________________________________________________
-
__________________________________________________________________________________________
-
__________________________________________________________________________________________
-
__________________________________________________________________________________________
-
__________________________________________________________________________________________
-
__________________________________________________________________________________________
-
__________________________________________________________________________________________
This will take you back if you came here by clicking a link: