 Morphological Image Processing
We have developed an algorithm to automatically segment 3-dimensional imagery.
The algorithm is designed to segment brain tissue conatined in
three-dimensional magnetic resonance imaging (MRI) datasets. The algorithm
fuses morphological filters by reconstruction (which analyzes the geometrical
information) with histogram-based thresholding (for gray level tissue
classification). Segmentation is performed by watershed analysis of the 3D data
set. The method effectively discriminates the brain tissue from the rest of the
anatomical structures within the MR signal volume in a fully automated fashion.
The robustness of this technique has been successfully tested on numerous
patient data sets.
Project Members:
References:
- "Automatic 3-dimensional segmentation of mr brain tissue using filters
by reconstruction,"
Joaquin Madrid and Norberto Ezquerra. In Petros
Maragos, Ronald Schafer, and Akmal Butt, editors, Mathematical Morphology
and its Applications to Image and Signal Processing, pages 417-424. ISMM,
Kluwer Academic Publishers, Boston, MASS, May 1996.
- "Topological Considerations on Grey Level Skeletonization,"
with
J. Madrid and R. Mersereau. Proc. Conf. on Visual Comm. and Image
Processing, SPIE V. 1818, p. 392-401, Boston, MA, 1993.
|
|
| | |
