The Application of Cellular Automaton in Medical Semiautomatic Segmentation - Cellular Automata in Image Processing and Geometry - page 168

Fig. 9.6 Two-labeling seeds in different MPR image planes

segmentation with few interactions as possible. The software is designed based on
object-oriented techniques (VC++ 8.0) to facilitate scheme improvements. The pro-
gram is implemented on a PC (Intel Dual Core E7200 (2.53GHZ) processor, 2GB
RAM, ATI Radeon HD 4850 graphics card with 512MB graphical memory), and
the GPU method is performed using OpenGL Shading Language (GLSL).
In MPR image planes, we only select a single axial slice to label desired tissues,
sagittal and coronal planes are similar to axial operations, Figure 9.8 demonstrates
the labels in axial plane. The total evolution time of implementation on the whole
MRI data takes about 812500 millisecond. Average computing time on each slice is
14.5 seconds. The experimental results are listed in Figure 9.9. Another important

a b c

Fig. 9.7 Samples of MRI Series. (a) Location=5. (b) Location =25. (c) Location =50.