Information Technology Reference
In-Depth Information
his/her therapeutic decisions than we could if we were using the traditional image
recognition paradigm.
6 Conclusions Concerning the Advanced Classification and
Cognitive Interpretation of CT Coronary Vessel
Visualizations
The research so far has shown that one of the hardest tasks leading to the com-
puter classification and then the semantic interpretation of medical visualisations
is to create suitable representations of the analysed structures and propose effec-
tive algorithms for reasoning about the nature of lesions found in these images.
Visualisations of coronary vascularisation are difficult for computers to analyse
due to the variety of projections of the arteries examined. The methods developed
made use of ETPL(k) graph grammars and the IE graphs they generate as formal-
isms modelling coronary vascularisation structures. Such solutions make it possi-
ble to detect all symptoms important from the point of view of diagnostics,
appearing as various types of stenoses of the coronary arteries of the heart. An im-
portant stage in achieving this goal is to construct the right syntactic/semantic
analysis procedure (i.e. a parser) which, when analysing graph representations
created for 3D reconstructions of coronary vascularisation, automatically provides
complete information, defining the 3D topology of the analysed graph that de-
scribes the coronary vessels including their individual components. Difficulties in
implementing a syntactic analyser are due to the lack of ready (i.e. software)
grammar compilers like those available for context-free grammars, for instance
[11][20]. This means that syntactic analysis procedures have to be derived inde-
pendently for the proposed grammars. A significant benefit of using ETPL(k)
graph grammars is the possibility of extending them to the form of probabilistic
grammars and the ability to introduce derivational rules with simple semantic ac-
tions [11][16]. This makes it possible, in addition, to determine significant mor-
phometric parameters of the analysed spatial reconstruction of coronary vessels.
Carrying out semantic actions assigned to individual productions generates certain
values or information which results from the syntactic analysis completed. In the
case of analyses of 3D coronary vascularisation reconstructions, semantic actions
of some productions will be aimed at determining numerical values defining the
location and degree of the stenosis as well as its type (concentric or eccentric).
These parameters will then be utilised as additional information useful in
recognising doubtful or ambiguous cases or symptoms. Even though every ana-
lysed image has a different representation describing its morphology, a syntactic
analysis is conducted based on defined grammars and this reduces the number of
potential cases to the recognition and inclusion of the examined structure in one of
the classes representing individual categories of disease units. If the completed
analysis does not end in recognising any pathological symptom, then the analysing
system generates information on the output that the analysed structure is a healthy
one. If symptoms which are ambiguous from the medical point of view are
Search WWH ::
Custom Search