From Individual to Population: Challenges in Medical Visualization - Scientific Visualization: Uncertainty, Multifield, Biomedical, and Scalable Visualization - page 249

there were published examples of computer-generated images, based on medical data
and used for medical applications.
During the past decades, medical image acquisition technology has undergone
continuous and rapid development. It is now possible to acquire much more complex
data than ever before. For example, in High Angular Resolution Diffusion Imaging
(HARDI), forty or more diffusion-weighted volumes are acquired in order to calcu-
late and visualize water diffusion and, indirectly, structural neural connections in the
brain [ 70 ]. In fMRI-based full brain connectivity, time-based correlation of neural
activity is indirectly measured between all pairs of voxels in the brain, thus giving
insight into the functional neural network [ 24 ]. Moreover, the questions that users
attempt to answer using medical visualization have also become significantly more
complex.
In this paper, we first give a high-level overview of medical visualization develop-
ment over the past 30 years, focusing on key developments and the trends that they
represent. During this discussion, we will refer to a number of key papers that we
have also arranged on the medical visualization research timeline shown in Fig. 23.1 .
Based on the overview and our observations of the field, we then identify and discuss
the medical visualization research challenges that we foresee for the coming decade.
23.2 Thirty-year Overview of Medical Visualization
Already in 1978, Sunguroff and Greenberg published their work on the visualization
of 3D surfaces from CT data for diagnosis, as well as a visual radiotherapy planning
system, also based on CT data [ 64 ]. Five years later, Vannier et al. published their
results developing a system for the computer-based pre-operative planning of cran-
iofacial surgery [ 71 ]. By this time, they had already used and evaluated their surgical
planning system in treating 200 patients. The system was based on the extraction
and visualization of 3D hard and soft tissue surfaces from CT data. Through the
integration of an industrial CAD application, it was also possible to perform detailed
3D measurements on the extracted surfaces.
23.2.1 Practical and Multi-modal Volume Visualization
In 1986, Hohne and Bernstein [ 26 ] proposed using the gray-level gradient to per-
form shading of surfaces rendered from 3D CT data. In 1987, Lorensen and Cline
published the now famous Marching Cubes isosurface extraction algorithm, which
enabled the fast and practical extraction of 3D isosurfaces from real-world medical
data. In the year thereafter, Levoy [ 47 ] introduced the idea of volume raycasting
in May, and Drebin et al. [ 19 ] in August. Although medical volume visualization
was possible before these publications, as witnessed by a number of publications,
previous techniques were either not as fast or yielded less convincing results. With