Anatomic VisualizeR's 3D Models
The 3D polygonal models which constitute the nucleus of Anatomic VisualizeR can be derived from a wide variety of sources. Current lessons are based, in large part, on anatomic polygonal models derived from the National Library of Medicine's Visible Human Project™ dataset. These models were produced and supplied by Visible Productions of Fort Collins, Colorado. Other models have been created by our in-house 3D modeler/medical illustrator.
Although the models supplied by Visible Productions were of high quality, implementation issues quickly emerged that had to be addressed with the end-user in mind. In order to elucidate particular learning objectives, careful consideration was given to determining the necessary level of anatomic accuracy of each model and the most appropriate model groupings.
Faculty often request specific modifications to change certain idiosyncratic structures in order to represent those seen in a higher perentage of the general population.
The impact of learning objective-driven choices on application performance was also an important consideration. For example, there is a finite practical limit to the number of polygons that can be displayed at one time, depending on technical limits, such as a particular machine's graphics hardware capability or whether its display is monoscopic or stereoscopic.
Taking both accuracy and performance issues into account, the combined polygonal count in each VR scene was reduced as much as practical without interfering with curricular objectives. Several approaches to decimation have been tried, and even after reductions of 70%-90%, plausible results were approved by expert anatomists. In addition to decimation, careful editing of facets and adjusting of polygonal patterns were required to optimize application performance.
The need to create new anatomical structures led to the use of alternative methods and approaches to 3D modeling. New models must be scaled and spatially registered with existing models. Template Graphic Software Inc.'s, Amapi 3D, Kinetix's 3D Studio Max and VP Sculpt are employed in these efforts by our in-house medical illustrator for missing and graphically problematic models.
In addition, 3D conceptual schematics needed for lessons are designed and developed by our in-house 3D modeler/medical illustrator in collaboration with clinical anatomists/subject matter experts.
SGI tools are employed to differentiate models from each other according to their color, shininess and transparency attributes. In the future, texture-mapping solutions will be used as a computationally cost-effective means to enhance visual realism.
