Some recent results in shape reconstruction
—
filed under:
Séminaire mensuel,
Séminaire
Séminaire mensuel du LIRIS par Dominique Attali, CR CNRS, Gipsa-lab
| What |
|
|---|---|
| When |
Apr 05, 2011 from 10:30 AM to 12:00 PM |
| Where | Amphi Claude Chappe, INSA de Lyon |
| Contact Name | S. Servigne et G. Damiand |
| Contact Phone | +33 (0)4.72.43.26.62 |
| Add event to calendar |
vCal
iCal
|
In many practical situations, the object of study is only known through a finite set of possibly noisy sample points. It is then desirable to try to recover the geometry and the topology of the object from this information. The most obvious example is probably surface reconstruction, where the points are measured on the surface of a real world object. In other applications, the shape of interest may live in a higher dimensional space, as for instance in machine learning and in particular in manifold learning. In this talk, we will review some recent results that account for the "good" behavior of some reconstruction algorithms observed in practice. These results typically formulate preconditions (or sampling conditions) under which the algorithm is guaranteed to output a topologically correct reconstruction.
