Introducing a mathematical model of noise in stereo images, we propose
a new criterion For intelligent statistical inference about the scene
we are viewing by using the geometric information criterion (geometri
c AIC). Using synthetic and real-image experiments, we demonstrate tha
r a robot can lest whether or not the object is located very far away
or the object is a planar surface without using any knowledge about th
e noise magnitude or any empirically adjustable thresholds.