Watersheds in Digital Spaces: An Efficient Algorithm Based on Immersion Simulations
IEEE Transactions on Pattern Analysis and Machine Intelligence
SIAM Review
| Hi-index | 0.00 |
We propose an efficient and non-parametric hierarchical partitioning approach that operates on a general 3D shape. The method suggested makes it possible to process an arbitrary shape provided that it can be described as a triangle mesh. The basic idea is to extend the use of the basin districts concept on curved spaces - such a partitioning process being applied on valuation issued from the computation of main curvatures over a polyhedral support. Hierarchical construction of basin districts is obtained from a Watershed Transform. In our working case, an efficient use of this well known transformation implies to be able to manage front propagations on a polyhedral surface with speed controled by the local characteristics of surface geometry. Moreover, the ability to provide an intrinsic shape partition from any triangle mesh lead us to propose a robust extension to usual mesh simplification algorithms.