Image Processing
Shape Warping and Statistics
In september 2003, in Taormina (Sicily), I took part in the workshop
« Designing Tomorrow's Category-Level 3D Object Recognition Systems: An
International Workshop », where I presented a poster on shape statistics,
here available (pdf or ps.gz, 900k).
Abstract:
Differentiable approximations of shape metrics (such as the Hausdorff
distance) provide an simple way to warp a shape onto another by solving a
Partial Differential Equation (PDE). Their first order variation defines a
normal deformation field for a given curve, from which we can define the
mean of several shapes, their covariance "operator", and the principal
modes of variation.