MR imaging contrast enhancement and segmentation using fuzzy clustering
One of the most important stages in medical image analysis is
objects segmentation. Segmentation results can be heavily affected
by image quality. Medical images usually present undesired
properties such as low signal-to-noise (SNR) and contrast-to-noise
(CNR) ratios, as well as multiple and discontinuous edges. This
explains why image enhancement takes an important role in the
segmentation and analysis process result. Our aim in this paper is to
present a method for medical image segmentation based on the
fuzzy c-means (FCM) algorithm preceded by a local image contrast
enhancement procedure. This method can be considered as a kind of
convolution filter but presents the originality of the adaptive found
of convolution mask coefficients. The grey level distribution of
pixels in the neighborhood of the current pixel is considered as 1/r2
distribution, which was deduced from the Newtonian model, where r
is a hybrid distance which involves the spatial information and the
luminance one. Finally, some results are presented in order to show
the computational performance of this approach.
Keywords: Image enhancement, Image segmentation, Magnetic resonance imaging, Fuzzy clustering
Download Full-Text
