Feature Extraction fro MRA I ages
for Fuzzy Rule-Based Diagnosis of Cerebral Aneurysms
Mika Ninomiya,Syoji Kobashi,Katsuya Kondo, Yutaka Hata, Seturo Imawakisup>* and Makoto Ishikawasup>*
Ishikawa Hospital
Proc. of 1st Int. Conf. on Soft Computing and Intelligent Systems,
CD-ROM, (2002)
Cerebral saccular aneurys often occurs at
bifurcations of blood vessels.This paper shows a ethod for
esti ating a fuzzy degree for high incidence rate of cerebral
aneurys .Our ethod consists of (1)seg enting the blood
vessels fro the MRA i ages,(2)skeletonization and detection of
the bifurcations,and (3)esti ating a fuzzy degree for high
incidence rate of cerebral aneurys .The esti ation ethod uses
fuzzy if-then rules representing physicians ’knowledge about the
incidence rate of the cerebral aneurys occurred at bifurcations.
The knowledge is about ‘location ’,‘area ’and ‘angle ’.
Experi ental results on seven normal healthy volunteers showed
that our ethod could give high fuzzy degrees for bifurcations of
the anterior com unicating artery (ACA).The detection rate of
the bifurcations of ACA thresholded by the fuzzy degree of 0.65
was 85.7%.