Image Processing Reference
In-Depth Information
questions that users attempt to answer using medical visualization have also become
significantly more complex.
This chapter first gives a brief overview of developments in medical visualization
over the past three decades. Basic techniques such as isosurface and direct vol-
ume rendering are discussed, as well as more advanced methods for multi-modal,
multi-field, multi-subject, and time-dependent data visualization. These techniques
are useful for therapy planning, predictive simulation, and diagnosis. In addition,
illustrative medical visualization is discussed, which is useful for presentation and
exploration purposes. Then, major medical visualization research challenges for the
coming decade are discussed. These arise first because of advances in hardware
and data acquisition, such as combined CT-PET scanners, simultaneous EEG-fMRI
acquisition, high angular resolution diffusion imaging (HARDI), molecular imaging,
high-resolution microscopy imaging, etc., leading to ever larger and more complex
data sets. Mobile display and computing devices also may have a great impact on
medical practice, leading to demands for new interaction and visualization for such
devices, and introducing tele-medicine. Other challenges are the interactive segmen-
tation of medical data, the integration of predictive simulation models and uncer-
tainty visualization in surgical planning, intelligent data mapping and reformatting,
and evaluation of illustrative versus (hyper)realistic visualization for diagnostic and
treatment planning purposes. Furthermore, visual analysis in healthcare as well as
visualization of population data are expected to grow in importance.
Chapter
24
is devoted to ultrasound imaging. Ultrasound is one of the most
frequently used imaging modalities in medicine due to its high spatial resolution,
interactive nature, and patient-friendliness. The main challenge of ultrasound is
image interpretation for diagnostic purposes, which requires extensive training.
Special problems arise because of the low dynamic range, noise and speckle occur-
ring in ultrasound images.
Ultrasound imaging presents several challenges for visualization. For example, in
functional ultrasonography, that is, ultrasound imaging of physiology and/or organ
function, information on motility, biomechanics and perfusion can be obtained non-
invasively. A set of 2D images can be aligned to form 3D data sets for which volume
visualization provides added value in terms of a more holistic understanding of
the data. Typical examples are demonstration of complex anatomy and pathology,
pre-operative surgical planning or virtual training of medical students. In addition,
matrix 3D probes are now on the market that allow real-time 3D acquisition. To ben-
efit from the high temporal resolution, advanced graphics techniques are required to
develop fast and efficient algorithms for rendering on-the-fly. Co-registration tech-
niques enable the use of multi-modal data sets. Fusion imaging, where ultrasound is
combined with either CT, MRI, or PET images, allows for more precise navigation
in ultrasound-guided interventions. This demands advanced visualization research
to enlighten how different data types can be combined and presented in novel ways.
This chapter presents the process-pipeline for ultrasound visualization, with an
overview of the specific tasks performed. A technique-based taxonomy is presented
based on a set of significant publications. In pre-processing, the ultrasound data is
reconstructed and oftentimes enhanced to improve quality. Segmentation techniques
Search WWH ::
Custom Search