Segmentation steps:
Figure 1: Input points selected by the user along the coronal axis.
Input points can be scattered around the surface and can be taken in any direction.
Figure 2: Three cross-sections of a RaG surface with sigma=3.0 and number of nodes=36 approximating the points in Figure 1.
A rational of Gaussian (RaG) surface is fitted to the input points to define the first estimate of the tumor boundary.
Figure 3: Edge points resulting from the nearest-neighbor search.
For every point on the boundary from Figure 2, an edge will be assigned from the zero-crossings image. By the nearest-neighbor approach, the closest edge point will be selected. Figure 3 shows three cross-sections along with resulting edges after the search. In order to provide a closed tumor surface, an elastic surface will be fitted to the edge points.
Figure 4: Cross-sections of the final result obtained by fitting a RaG surface to the edge points in Figure 3.
Figure 4 shows three slices of the segmentation result after the RaG surface is fitted to the edge points.
[Go Back] [CSE Department Home Page] [Intelligent Systems Lab Page]
For more information contact: A. Goshtasby (ardeshir@cs.wright.edu).
Last modified: 8/21/97.