Information Technology Reference
In-Depth Information
recognised, the final diagnosis can only be made by applying additional analysing
algorithms.
The development of new methods of computer analysis and detection of
stenoses inside coronary arteries helps not only to significantly improve diagnostic
actions, but also greatly broadens the application spectrum of artificial intelligence
in computer understanding of diagnostic images and determining the medical sig-
nificance of pathologies shown in them. The linguistic formalisms developed also
add new types of grammars and their applications to the fields of artificial intelli-
gence and image recognition. Such techniques are of major importance as they al-
low lesions not just to be recognised, but also their semantics defined, which in the
case of medical diagnostic images can lead to the computer understanding their
significance. This is of key importance for detailing the best therapeutic possibili-
ties and if the proposed methods are perfected, they can significantly improve the
ability to support the early recognition and diagnostics of heart lesions. This is of
practical significance as the identification of locations of stenoses in coronary ves-
sels is performed very widely, but manually, by an operator or a diagnostician, and
as the research has shown, such key stages of the diagnostic process based on the
analysis of 3D images can, in the future, be successfully executed by the appropri-
ately designed computer system. It is also worth noting that the methods presented
in this publication are not just an attempt at assigning the examined image to an
a'priori defined class, but are also an attempt to imitate and automate the human
process of the medical understanding of the significance of a shape found in the
analysed image. This approach allows numerous medical conclusions to be drawn
from the examined image, which, in particular, can lead to making the right diag-
nosis and recommending a specific type of therapy depending on the shape and
location of the pathological stenosis described in the sets of productions. There is
a deep analogy between the operation of the structural analysis model and the
cognitive interpreting mechanisms occurring in the human mind. The analogy
consists in using the interference between the expectations (knowledge collected
in the set of productions in the form of graph grammatical rules) and the stream of
data coming from the system analysing the examined image. This interference is
characteristic for the human visual perception model [20].
Problems related to automating the process of generating new grammars for
cases not included in the present language remain unsolved in the on-going re-
search. However, it is worth noting that generally, the problem of deriving
grammatical rules is considered unsolvable, particularly for graph grammars. It
can appear if the image undergoing the analysis shows a coronary vascularisation
structure different from the so far assumed three cases of vessel topologies
occurring the most often, i.e. the balanced distribution of arteries, the dominant
right artery or the dominant left artery. In those cases it will be necessary to define
a grammar taking this new case into account. The processes of creating new
grammars and enriching existing ones with new description rules will be followed
in further directions of research on the presented methods. Another planned ele-
ment of further research is to focus on using linguistic artificial intelligence
method to create additional, effective mechanisms which can be used for indexing
and quickly finding specialised image data in medical databases. Such searches
Search WWH ::
Custom Search