Summary: Possibilistic Hopfield Neural Network on CT Brain Hemorrhage Image
Dachuan Cheng, *Qin Pu, Kuosheng Cheng, and **Hans Burkhardt
Institute of Biomedical Engineering, National Cheng Kung University,
Tainan, Taiwan, ROC.
*Lab. BI, Internal Medicine, Freiburg University Hospital, 79110, Freiburg, Germany.
**Informatik Institut, Freiburg Universität, 79110, Freiburg, Germany.
In this paper, a possibilistic Hopfield neural network
(PHNN) has been proposed for clustering and subsequently
applied to brain hemorrhage image segmentation based on
a series of CT images. The neural network structure has
been implemented by imbedding the weighting possibilistic
cmeans algorithm into a Hopfield neural network. The
network solved the coincidental cluster problem by using a
weighting factor and it can also be implemented in parallel.
The proposed neural network has been compared to fuzzy
cmeans (FCM), possibilistic cmeans (PCM), and fuzzy
possibilistic cmeans (FPCM) algorithms by using both
simulated data and real images. The results showed that