Abstract:
|
The purpose of this report is to analyze the current methods of data
integration in Computer-Assisted Neurosurgery applications.
Neurological studies require the integration of anatomical data
registered with CT Computed Tomography, MRI Magnetic Resonance and MRA
Magnetic Resonance Angiography with functional data from Nuclear
Tomographies, PET and SPECT along with fMRI Functional Magnetic
Resonance and MEG Magnetoencelography.
The first section of the document presents the physical model
corresponding to the human brain, it describes the anatomy, the
physiology and the pathologies that need the assistance of imaging
modalities. The second section analyzes the type of data provided by
the devices above enumerated, their values, type and range, the
physical properties that they measure and their resolution. The impact
of the registration on the patients is also evaluated taking into
account the invasivity of the technique and its harmfulness. The third
and the fourth sections focus at the integration itself. First, a
general overview of the integration process is given. Next, different
integration strategies are presented and compared. The fifth section
addresses the modelling of integrated data. In the sixth section, the
error of each step of the integration is analyzed. Finally, several
strategies for the joint visualization of several data types are
discussed. In the conclusions, a global evaluation of the different
methods is presented, outlining their lacks and their advantages. |