The gray level values are not used in the connectivity
calculation. Instead, they are used in classifying the tissues. Therefore, 3D
volume data can be divided into two types; values above and below a surface
value that defines the surface of interest. The connectivity algorithm extracts
an iso-intensity surface by examining each slice from the three dimensional
data matrix.
This algorithm needs manual selection of a surface value. Then it
evaluates each point in the 3D lattice with respect to the chosen surface
value. Points in the volume matrix having less intensity than the selected
surface value are replaced by a zero; while points having intensity greater
than or equal to the surface value are replaced by a one. The surface points
then can be determined by examining the six neighbors of each point. A point
that has at least one zero neighbor is by definition on a surface, and points
surrounded by ones are internal to the surface.
The extraction of a selected surface requires the manual selection of a ``seed'' point that is internal to the surface of interest. The algorithm then identifies all points that are connected to the seed point and are internal to the surface. Two points are connected if there exists a path between adjacent lattice points that are of the same value. Therefore each point is connected to six adjacent points, represented by four adjacent pixels in each slice and two pixels at the same location in adjacent slices. By using a recursive procedure, we start with the seed point and mark each of the six adjacent unmarked points that are of the same type as the seed point. These connected points become the new seed points for the next depth recursion. The procedure continues for all active points that are tagged by negating the data point intensity. During surface reconstruction step connectivity algorithm only processes tagged points.