From BlenderWiki

Shape Enhanced

Shape enhanced method exaggerates a shape using a Laplacian smoothing operator in the reverse direction.

Methodology

A diffusion process is used:

δV / δt = λL(V)

Integrating the diffusion equation with a explicit Euler scheme:

(I − | λdt | W_pL)V' = V^t

V^{t + 1} = V^t + sing(λ)(V^' − V^t)

The vertices V^{t + 1} are enhanced in the reverse direction of curvature flow.

Where L is the Laplace Beltrami Operator, V' are the smoothing vertices, V^t are the actual vertices positions, W_p is a diagonal matrix with weight vertex group, and λdt is the enhancement factor that supports negative and positive values: negative for enhancement and positive for smoothing.

Implementation

modifiers/intern/MOD_laplaciansmooth.c

Implementation of this method only require small changes in Laplacian smooth method, the changes are presented in this patch.

Is necessary adding a checkbox in Laplacian Smooth Modifier interface, to customize the kind of Laplace operator, (Traditional or Normalized).

Results

Top row: From left to rigth. Shape enhanced with λ = − 400.0 , λ = − 100.0, Original model, λ = 50.0. Bottom row: Shape enhancement with weight vertex group, λ = − 50.0 for the legs, λ = − 200.0 for the head and neck.