|
|
We have conducted two experiments, one with two plastic pipe simulations generated at the University Medical Center of SUNY Stony Brook, and the other using the Visible Human data set. The objective of the first study is to see how well we can detect the presence and characteristics of mucosal lesions measuring at least one centimeter in diameter, which are considered clinically significant since they have a high probability of being malignant. As prototypes and for testing purposes, two different phantoms consisting of a plastic pipe looped inside a water tank were imaged by a GE HighSpeed CT in the helical mode, which resulted, respectively, in one 512x512x280 and another 512x512x107 volumetric data set, with each sample point represented in 2 bytes. The plastic pipe had a radius of 2cm, and its surface had corrugated and smooth segments. For the 512x512x280 a rubber cylinder of length 7mm and diameter 5mm was inserted into the pipe to simulate a tumor. This data set was obtained with a slice thickness of 3mm at increments of 1mm using a field-of-view of 34.5cm.
Figure 8 shows six images in a virtual interactive flythrough of the 512x512x280 pipe data set, generated using VolVis. Figure 8a is a view of the phantom before entering the pipe; Figure 8b shows the image of the pipe before the tumor is encountered; Figure 8c is a distant view of the tumor within the pipe before reaching it; Figure 8d is a close front view of the tumor (notice that the viewing orientation is different from that of Figure 8c); Figure 8e is a close side view of the tumor when passing next to it; and Figure 8f is a distant view of the tumor after passing it.
For the 512x512x107 data set, three rubber cylinders were inserted into the pipe. The lengths of the three cylinders were 7mm, 5mm, and 3mm, respectively, and the diameters were 5mm for all cylinders. This data set was obtained with a slice thickness of 5mm at increments of 2.5mm using a field-of-view of 40cm. A flythrough animation of the 512x512x107 pipe data set has been generated automatically with the technique described in Section 5 and recorded on a video tape, which is included in the video proceedings.
In our second study, we successfully applied the automatic animation technique of Section 5 to the Visible Human data set. In this study, we primarily focused on the physical cross-sections of the data set, from slice No. 1595 to slice No. 1848. Due to memory limitations, each slice of the physical cross-sections was down-sampled from the original resolution of 2048x1216 to 683x406. The red component of the 24-bit color data set was extracted to compute the eye positions and viewing directions of the flight path along the colon of the Visible Human, and later, the whole 24-bit color data set was used as a texture map when the animation frames were rendered by VolVis.
Figure 9 shows six frames of this flythrough animation. Figure 9a is the start of flight through the descending colon proceeding towards the rectum; Figure 9b shows the colon curving to the left; Figure 9c shows three haustra (inner-folds) of the colon; Figure 9d is a close view of the three haustra seen in Figure 9c; Figure 9e shows the colon beyond the last haustra seen in Figure 9c; and Figure 9f shows a smooth walled object projecting from the colon surface. Compared with the images in Figures 9a through 9e, the large size of the lumen in Figure 9f indicates that we are entering the rectum. The complete flythrough animation is also included in the video proceedings.
