Virtual Cadaver – Supercomputing
The Virtual Cadaver is a visualization of data provided by the Visible Human Project of the National Library of Medicine. This project aimed to create a digital image dataset of complete human male and female cadavers.
The male dataset originate from Joseph Paul Jernigan, a 38-year-old convicted Texas murderer who was executed by lethal injection. He donated his body for scientific research in 1993. The female cadaver remains anonymous, and has been described as a 59-year-old Maryland housewife who passed away from a heart attack. Her specimen contains several pathologies, including cardiovascular disease and diverticulitis.
Both cadavers were encased in gelatin and water mixture and frozen to produce the fine slices that comprise the data. The male dataset consists of 1,871 slices produced at 1 millimeter intervals. The female dataset is comprised of 5,164 slices.
The Duderstadt Center was directed to the dataset for the female subject in December of 2013. To load the data into the virtual reality MIDEN (a fully-immersive multi-screen head-tracked CAVE environment) and a variety of other display environments, the images were pre-processed into JPEGs at 1024×608 pixels. Every tenth slice is loaded, allowing the figure to be formed out of 517 slices at 3.3mm spacing per slice. A generic image-stack loader was written to allow for a 3D volume model to be produced from any stack of images, not specific to the Visible Human data. In this way, it can be configured to load a denser sample of slices over a shorter range should a subset of the model need to be viewed in higher detail.
Users can navigate around their data in passive cinema-style stereoscopic projection. In the case of the Virtual Cadaver, the body appears just as it would to a surgeon, revealing the various bones, organs and tissues. Using a game controller, users can arbitrarily position sectional planes to view a cross-section of the subject. This allows for cuts to be made that would otherwise be very difficult to produce in a traditional anatomy lab. The system can accommodate markerless motion-tracking through devices like the Microsoft Kinect and can also allow for multiple simultaneous users interacting with a shared scene from remote locations across a network.