Transforming clinical imaging data for virtual reality learning objects

Abstract

Advances in anatomical informatics, three-dimensional (3D) modeling, and virtual reality (VR) methods have made computer-based structural visualization a practical tool for education. In this article, the authors describe streamlined methods for producing VR "learning objects," standardized interactive software modules for anatomical sciences education, from newer high-resolution clinical imaging systems data. The key program is OsiriX, a free radiological image processing workstation software capable of directly reformatting and rendering volumetric 3D images. The transformed image arrays are then directly loaded into a commercial VR program to produce a variety of learning objects. Multiple types or "dimensions" of anatomical information can be embedded in these objects to provide different kinds of functions, including interactive atlases, examination questions, and complex, multistructure presentations. The use of clinical imaging data and workstation software speeds up the production of VR simulations, compared with reconstruction-based modeling from segmented cadaver cross-sections, while providing useful examples of normal structural variation and pathological anatomy.