Affiliations: Department of Electronics and Communication Engineering, Mepco Schlenk Engineering College, Sivakasi, Tamilnadu, India
Correspondence:
[*]
Corresponding author: D. Selvathi, Department of Electronics and Communication Engineering, Mepco Schlenk Engineering College, Sivakasi, Tamilnadu, India. E-mail:[email protected]
Abstract: Alzheimer's disease (AD) is the most common type of dementia and a major
cause of disability worldwide.It is a progressive and degenerative disease
that affects brain cells and its early diagnosis has been essential for
appropriate intervention by health professionals. In this work, Random
forest classifier is used to identify Alzheimer's disease (AD) affected
patient using brain magnetic resonance imaging (MRI). Random Forest grows
many classification trees. To classify a new input vector, each tree votes
for a class and the forest chooses the class having the majority of votes
over all the trees in the forest. A set of MR brain images from healthy and
pathological subjects extracted from the OASIS (Open Access Series of
Imaging Studies) database are used to evaluate the performance of the
proposed method. Nonsubsampled Contourlet Transform decomposes the input
image into many sub bands. Features are extracted from the input images and
also from the transformed images. In this work, two groups of features are
used to extract the information. In the first groups, features such as
intensity, orientation, and edge information & in the second group,
features such as color Mean, Symmetry, Center Black area, Total Brain Area,
Mean, Variance, Skewness, Image Kurtosis, Gradient Mean and Gradient
Variance are calculated. The features are given as the input to Random
forest classifier and also to the Support Vector Machine (SVM) for comparing
the performance of proposed work with the well-known existing SVM
classifier. From the classification results, it is observed that Random
forest classifier performed better than the SVM classifier and other
existing work.
Keywords: Alzheimer's disease, SVM, random forest classifier, magnetic resonance imaging, NSCT
