RELIEFF FEATURE SELECTION BASED ALZHEIMER DISEASE CLASSIFICATION USING HYBRID FEATURES AND SUPPORT VECTOR MACHINE IN MAGNETIC RESONANCE IMAGING
Journal Title: International Journal of Computer Engineering & Technology (IJCET) - Year 2019, Vol 10, Issue 1
Abstract
Alzheimer disease is a form of dementia that results in memory-related problems in human beings. An accurate detection and classification of Alzheimer disease and its stages plays a crucial role in human health monitoring system. In this research paper, Alzheimer disease classification was assessed by Alzheimer’s disease Neuro-Imaging Initiative (ADNI) dataset. After performing histogram equalization and skull removal of the collected brain images, segmentation was carried-out using Fuzzy C-Means (FCM) for segmenting the white matter, Cerebro-Spinal Fluid (CSF), and grey matter from the pre-processed brain images. Then, hybrid feature extraction (Histogram of Oriented Gradients (HOG), Local Binary Patterns (LBP), and Gray-Level CoOccurrence Matrix (GLCM)) was performed for extracting the feature values from the segmented brain images. After hybrid feature extraction, reliefF feature selection was used for selecting the optimal feature subsets or to reject the irrelevant feature vectors. Then, the selected optimal feature vectors were given as the input to a supervised classifier Support Vector Machine (SVM) to classify three Alzheimer classes of subjects; those are normal, Alzheimer disease and Mild Cognitive Impairment (MCI). The experimental outcome showed that the proposed methodology performed effectively by means of sensitivity, accuracy, specificity, and f-score. The proposed methodology enhanced the classification accuracy up to 2-20% compared to the existing methodologies.
Authors and Affiliations
HALEBEDU SUBBARAYA SURESHA and S. S. PARTHASARATHI
Keywords
Related Articles
- EP ID EP46516
- DOI 10.34218/IJCET.10.1.2019.015
- Views 238
- Downloads 0
RESEARCH TRENDS IN HYPERSPECTRAL IMAGERY DATA
Remote sensing and analysis of an object or specific area of the earth at different distances with very large number of bands forms a major part of hyperspectral imaging technology. Currently, a wide range of data sets...
AN OPTIMAL COMPOSITION PLAN SELECTION USING MULTI OBJECTIVE PARTICLE SWARM OPTIMIZATION
Domain-ontology based Particle Swarm Optimization (PSO)-inspired Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH) and Improved Bipartite graph is an efficient web service composition approach. It co...
IMPLEMENTATION OF STABLE PRIVATE CLOUD USING OPENSTACK WITH VIRTUAL MACHINE RESULTS
In today’s era educational organization strongly needs devices which are ready to access and use and also various operating system platforms are required for different learning courses. To achieve this type of environm...
SECURE DATA TRANSMISSION THROUGH NODE-DISJOINT ON DEMAND MULTIPATH ROUTING IN MANETS
Mobile Ad Hoc Networks (MANETs) are the wireless networks which can be deployed instantly without requiring any fixed wired infrastructure. MANETs are specifically very much useful in military, commercial and civilian...
SECURED DATA AGGREGATION USING FIBONACCI NUMBERS AND UNICODE SYMBOLS FOR WSN
Wireless Sensor Network is a combination of one and more nodes basically used for data collection and data aggregation. The collected data is aggregated through a data aggregator and it will be sent to bases station. W...