Next: References Up: Title Page Previous: Slicing Considerations Index: Full Text Index Contents: Conference Page

Results and Conclusions

The relationships between the three Visible Human Data Sets; National Library of Medicine (NLM), Spitzer, and Gold Standard are depicted in figure1 and table 1. Figure 1 specifically highlights the file sizes and orientation through the same slice (File) number from each Data Set. The most important attribute of these Data Sets is the orientation. The orientation denotes the correct view of the data for various medical professions. For example, the correct view for a radiologist is the Gold Standard view. It portrays the body with back down on the table from feet up to head. In table 1, the VH Data Set Chart outlines the various nuances and attributes of the three data sets in table form. It summarizes the original file format; conversion programs used on three different platforms PC, Mac, and Unix (SGI; file sizes and additional notes possibly pertinent to using the individual data set to it fullest potential.

The importance of registration or alignment is emphasized in figure 2 and figure 3. Our studies show that the NLM Data Set has certain offsets that need to be accounted for when creating 3D models. The methodology for comparing two (2) sequential slices/files is simple. If an offset is suspected, using Photoshop we highlighted the bottom left fiducial rod with a red or yellow circle on each file. Then set a corresponding red or yellow crosshair in the same position on both files. In one file we overlayed the crosshairs and the circles. The crosshairs should line up on top of one another and in practicality the circles should too, however, on files with offsets the circles are misaligned.

Specifically, figure 2 is an example of aligned sequential files, from the thoracic region, while figure 3 is an example of files with an offset. The last example, figure 4, shows a 3D model of aorta arch with offsets and a corrected 3D model with no offsets.

The Visible Human data is meant to be shared by those who are engaged in serious visualization of human anatomy. We wish to have a standard way (or a number of ways) of presenting not only the original raw data but also the outcome of the 3D visualization/reconstruction process. We must understand very well the formats and the content of the currently available variation of the data. We wish to have an open discussion between researchers and developers on establishing
standards which will help us to share and compare our work related to the Visible Human project.

The VH data sets continue to serve as a visual communication language for many fields of medicine [1]. Its application potential is only limited by imagination of the developers. Giving these developers the means to understand, interpret, and utilize the data sets to their fullest potential is the key to a successful application.

Next: References Up: Title Page Previous: Slicing Considerations Index: Full Text Index Contents: Conference Page