Image Processing Reference
In-Depth Information
the process should be automated, thus improving segmentation
throughput and repeatability. How to devise such efficient and intuitive
mechanisms for 3-D manipulation of models and volumetric data, and
how to integrate them into the deformation of the models, remain topics
for future research.
3.
Research on functional cardiac descriptors: There are many shape and
motion parameters other than traditional indices (cf.
Appendix A
).
Unfortunately, although these new indices seem to provide either richer
information or a more detailed analysis of cardiac function, their clin-
ical evaluation has been very limited. As a consequence, it is difficult
to determine their clinical relevance and the extra information provided
with respect to traditional indices such as LVV, EF, etc. The lack of
clinical evaluations may be related to the fact that advanced 3-D mod-
eling techniques, from which these parameters can be derived, are
computationally expensive and require considerable user intervention.
The need of considerable pre- and postprocessing procedures,
ad hoc
parameter settings, and technical understanding of the modeling tech-
nique itself may explain why most of the described approaches are not
available as stand-alone prototypes on which clinical studies can be
carried out routinely.
There is certainly room for development of novel shape and motion
descriptors. However, there is an even larger need for evaluation of
already-existing indices on reference data sets and large-scale clinical
studies. It is remarkable that this lack of large-scale evaluation studies
is present even in the case of many techniques aiming at the extraction
of traditional functional parameters (
Table 9.2
).
It is unrealistic to expect that every new technique proposed in the future
will go through the process of a thorough clinical evaluation study.
Unfortunately, many research institutes working on geometric modeling
and shape analysis are not located in a clinical environment. Access to
state-of-the-art image material and derived parameters for testing and
benchmarking purposes is, therefore, difficult. In this respect, a public,
common database of a representative set of images from different modal-
ities would be highly beneficial. This database should establish a few
standard data sets (both synthetic and clinical study cases) with as many
independent measurements as possible of mass, SV, etc. Given the cur-
rent speed of development in imaging modalities, such a database should
be updated regularly to be representative of the state-of-the-art imaging
technology.
4.
Research into image-based in silico cardiac modeling and simulation:
In recent years there has been a growing interest in research bridging
the gap across different structural and functional levels in computa-
tional modeling and simulation of the cardiac system [256-260].
This field requires integration of research findings from traditional
Search WWH ::
Custom Search