| Summary: | In medicine, there are diseases such as Alzheimer's are not yet found a way
of healing. But one thing is believed to be the cause of brain degeneration or reduced
levels of human brain tissue. Diagnosis is expected to continue to find ways to
prevent or slow down the degeneration of brain distribution.
The data used in the form of MRI, which requires a corresponding image
processing techniques to detect, locate and quantify the lost tissue in the early
commencement of the disease. Levels of brain tissue loss that can distinguish the
normal brain. The factors underlying the development of the program of
determining the location and quantification of brain degeneration with Gaussian
Mixture Model method. In the present study, the authors conducted immunity
testing method of Gaussian Mixture Models to the noise on the program that has
been developed previously in order to test the feasibility of a program to be realized
to be used in the medical community. The method has three stages, these are the
stage of Gaussian Mixture Models, Hidden Gaussian Mixture Models, as well as
localization and quantification. How the test is to provide a noise in the images to
be processed and executed on the method, and compare the MSE degeneration ROI
in the image without noise and image noise are given. Noise used in this study is
the noise that often arise in medical images. Because of the problems that may occur
is the appearance of noise in the image, thereby disrupting the process to be
executed
The conclusion of this study, which tested the method does not have
immunity to noise, after comparing the ROI degeneration MSE on images without
noise and noises that turned out to have much. While the comparison of the MSE
on the results of each stage of the method shows that Gaussian Mixture Models
phases have the weakest immune.
|
|---|