Haussecker, H.W. and Fleet, D.J.
Estimating optical flow with physical models of brightness variation.
IEEE Transactions on Pattern Analysis and Machine Intelligence,
Although brightness constancy is commonly assumed by optical flow
techniques, it is well-known that this produces poor estimates of
image motion under many conditions. This paper describes a generalized
formulation of optical flow estimation based on models of brightness
variation that have time-dependent physical causes. These include
changing surface orientation with respect to a directional illuminant,
motion of the illuminant, and physical models of heat transport in
infrared images. With these models we simultaneously estimate the
optical flow and the relevant physical parameters. The estimation problem
is formulated using total least squares (TLS), with confidence bounds
on the parameters. Experiments are shown in four domains with both
synthetic and natural inputs.
Return to David Fleet's home page.