| | |
Summary: Independent3D MotionDetectionBased on
Depth Eliminationin Normal Flow Fields
Antonis A. Argyrost, $ SteliosC. Orphanoudakist
tCVRL,/ICS/FORTH $CVAP/NADA/KTH
POBox 1385Heraklion, Crete, GR-711-10, Greece Stockholm,S-100 44, Sweden
E-mail: {argyros,orphanou}@ics.forth.gr
Abstract
Thispaper considers a spec@ problem of visualpercep-
tion of motion, namely the problem of visual detection of
independent 3 0 motion. Most ofthe existing techniquesfor
solving this problem rely on restrictive assumptions about
the environment,the observer's motion, or both. Moreovel;
they are based on the computation of optical jow, which
amounts to solving the ill-posed correspondenceproblem.
In this work, independent motion detection isformulated as
robust parameter estimation applied to the visual input ac-
quired by a binoculal; rigidly moving observer Depth and
motion measurements are combined in a linear model. The
parameters of this model are related to the parameters of
self-motion (egomotion)and the parameters of the stereo-
|
