| Summary: | 碩士 === 中原大學 === 資訊工程研究所 === 83 === In this thesis, we developed a data structure to represent
boundaries of interested tissues in medical volume. For each
voxel, we use a boundary pointer that points the next boundary
of any kind of tissue in depth direction and a pointer to its
reverse direction. By manipulating the boundary pointer data in
voxel, we can simulate the operation for the hard tissue and
the soft tissue that are associated with the hard tissue. We
interpolate the correct position between two voxels by their
values. Besides, using 3×3 sampling points that are determined
from 2×3×3 neighboring voxels, we can obtain a quadratic
approximated tissue surface on the voxel. Based the local
quadric surfaces, our system can provide high quality images.
We normalized the voxel value of every kind of tissue. That
means we can move the sectioned anatomical structure to any
place (of othervoxels) while need not to change the voxel
values of the structure. By manipulating the boundary position
data, we have developed the algorithms that allow the clinician
to section the anatomical structure from various directions and
move the sectioned structure to any place in spite of our
2(1/2)D boundary representation. We compute the new position of
soft tissue simultaneously with the reposition of the hard
tissue. We also provide the interactive environment to assist
the sophisticated operations.
|
|---|
