Summary: Deformable Template Methods for
Yali Amit 1
Department of Statistics
University of Chicago
November 5, 1998
1 Supported in part by the Army Research Office under grant DAAH049610061 and
MURI grant DAAH049610445,
Numerous computer vision problems consist of finding a certain object in an image.
Find the face in a photograph, find the road in a satellite image, find the ventricles in
an axial MRI image of the brain. Often it is not guaranteed that the object is present
in the image, or there may be more than one instance. The notion of `find' is also
very vague. One could require a precise outline of the object in the image, one could
require the detection of a certain number of well defined landmarks on the object.
One could define `finding' as identifying a deformation from a model or prototype of
the object into the image. The deformation could be a simple 2d affine map or a high
dimensional warp. The object itself may have different degrees of variability. It may