LIBSIMPLEX : Deformable Models Library for Image and Surface Reconstruction
This library includes more than 150 000 lines of C code and covers the different aspects of deformable surface modeling. The main objective of this library is to support the deformation of simplex mesh surfaces[1] for the surface reconstruction from a variety of 3D datasets:
- Clouds of points
: a set of unstructured data points
- Structured range images
: a 2D array of 3D points either in cartesian or cylindrical geometry
- Volumetric images
: 3D images stored in the Inrimage image format (for instance CT-scan or MR images) with associated information (gradient, direction of the gradient,...)
The library is structured into a set of modules that are dedicated to a given geometric data structures. The list of these modules is :
- Core module
: creates basics data structures (dynamic array, KD-tree,...)
- Image module
: I/O routines for volumetric images and basic image processing.
- Mesh module
: I/O routines for 3D triangulation and general mesh (may not be a manifold). Includes a powerful decimation algorithm
- Range image module
: I/O routine for range images (Cyberware or general purpose) and unstructured cloud of points. Include also a closest point algorithm for each dataset.
- Isosurface module
: Create isosurfaces and isocontours from volumetric images based on a modified marching cubes algorithm. Handles cutting and suturing of isosurfaces.
- Simplex mesh module
: Support the deformation of simplex meshes (and contours defined on simplex meshes) with or without external range or volumetric images. Also support the edition of mesh topology.
Two applications sm and yav, repectively based on Motif and Tcl/TK, are also available with this library. They include a complete 3D interface that allow to control interactively the deformation of simplex meshes.
|

Surface Segmentation |

Surface Reconstruction
|

Surface Segmentation |
Référence :
[1] H. Delingette. General object reconstruction based on simplex meshes. International Journal of Computer Vision, 32(2):111-146, September 1999. (PostScript)