You are on page 1of 6

Intracranial hemorrhage Arya

Sanya

detection using spatial


fuzzy c-mean and region-
based active contour on
brain CT imaging

H. S. Bhadauria · M. L. Dewal
Spatial An image can be represented in various feature
spaces and FCM segments the image into objects by
fuzzy c- grouping the similar pixels in the feature space.

means FCM clustering is an unsupervised clustering


technique applied to segment images into clusters
clustering with similar spectral properties. It utilizes the distance
between pixels and cluster centers in the spectral
domain to compute the membership function.
The clustering is achieved by iterative minimization
of a cost function that depends on the distance of the
pixels to the cluster centers in the feature domain.
Region Active contour is the method of segmentation
based active which uses energy forces to separate pixels.

contour This uses internal and external energy forces to


form a closed contours. External energy controls
model the location of contour whereas Internal energy
governs deformable changes.
Method

The method first segments the hemorrhagic area using


fuzzy clustering.The results of fuzzy clustering are used
to initialize the level set function as well as to estimate
the controlling parameters, which control the evolution of
level set function. Let Ui,j = [ui,j] be the membership
function for the fuzzy segmentation. The object of
interest from the given image can be extracted as binary
image Zi,j
Why spatial By incorporating region-based information into the energy

fuzzy c - function, this method has much convergence range and


flexible initialization.
means The standard FCM does not consider any spatial
clustering? information in image context, it suffers from the
disadvantage of being sensitive to noise and other
imaging artifacts. Recently, many researchers have
incorporated local spatial information into the standard
WHY region FCM.. The spatial function is the summation of the

based over membership function in the neighborhood of each pixel


under consideration.

edge Region-based models are not sensitive to objects with poorly

based ? defined boundaries but are sensitive to inhomogeneity of


image intensities, i.e., the overlapping of the intensity ranges.
Disadvantag Disadvantage : The way in which spatial fcrm

es incorporates the neighbouring info limits their


application to single feature input.

It suffers from the weakness such as initialization of


cluster centers and determining optimal no. of clusters.

You might also like