Browse all publications by topic
Browse all publications by year
- T. Otto, M. Scholz,
G. Galizia, and K. Obermayer. Automatic Reconstruction of Olfactory
Glomeruli from Confocal Image Stacks.
.
In European J. Neurosci., volume 10, page 413. European Forum of
Neuroscience ENA, 1998.
We present a semi-automatic method for extracting the three
dimensional arrangement of glomeruli from the honey bees antennal lobe out
of raw confocal microscope data. Current this process involves tedious manual
work, namely drawing the outlines of each glomerulus in every single confocal
frame by hand. Because the data is very noisy, threshold based surface
rendering methods like the well known marching cubes algorithm fail. Those
methods produce polygonal surfaces which consist of some 100,000 triangles,
by fitting the noise rather than providing a clean segmentation of the
glomerulis surfaces. In order to overcome these difficulties we have
developed a fast and semi-automatic method for extracting the three
dimensional shape of the glomeruli in the antennal lobe based on a geometric
method. Starting from a point clearly inside the glomerulus which should be
reconstructed, we cast equally spaced rays through the volume. By analyzing
the signal along these rays ``support points points are computed, which
lie on the glomerulus surface and which can then be used to generate the
triangulation. The method is fast, generates much less triangles than
conventional surface rendering methods, and compares very well with handmade
reconstructions. It is well suited for comparative studies.
|
