Computer Graphics Forum (Proceedings of Pacific Graphics 2010), accepted
L. Liu[1], E. Chambers[2], D. Letscher[2], T. Ju[1]
[1] Washington University in St. Louis, USA
[2] St. Louis University, USA


Thinning is a commonly used approach for computing skeleton descriptors. Traditional thinning algorithms often have a simple, iterative structure, yet producing skeletons that are overly sensitive to boundary perturbations. We present a novel thinning algorithm, operating on objects represented as cell complexes, that preserves the simplicity of typical thinning algorithms but generates skeletons that more robustly capture global shape features. Our key insight is formulating a skeleton significance measure, called medial persistence, which identify skeleton geometry at various dimensions (e.g., curves or surfaces) that represent object parts with different anisotropic elongations (e.g., tubes or plates). The measure is generally defined in any dimensions, and can be easily computed using a single thinning pass. Guided by medial persistence, our algorithm produces a family of topology and shape preserving skeletons whose shape and composition can be flexible controlled by desired level of medial persistence.

A gallery of input model and skeletons computed by our method in different dimensions. 1st row shows the input; 2nd & 3rd rows show our medial persistence measure on medial faces and edges respecitively; last two show the mixed skeleton and curve skeleton result with different thresholds.



Below shows how easy it is to use the program
Compute skeleton directly from mesh(.ply):
Step 1: load in a mesh.

Optional: Drag mouse, with left button pressed to rotate, middle button pressed to scale, and right button pressed to translate.

The mesh is automatically turned into octree represenation and cell complex representation, and saved into files. The process is slow. However, after this process, in the future, the model can be loaded in by reading the cell complex file, which is much faster.

Step 2: : click 2PassThinning to get the measure and the skeletons.

Face (left to right): Mabs, Mrel, and surface patch size
Edge (left to right): Mabs, Mrel, and curve length
Thresholds for Mabs is in the range [0, infinity), and for Mrel is [0,1]. If the thresholds for face is extremely large, for example, 1 for Mrel of face, then all faces will be removed during skeleton computation unless they are critical to maintian topology such as the surface of a hollow sphere.

Step 3: : Save result.

Visualization options

Tetrahedralize a model: the tetrahedralizaiton process is slow, be patient

The best way to learn is trying it out. Enjoy! :)
Lu Liu, 09/17/2010 12:03AM