Research Area: Machine Learning
Deep learning methods have gained more popularity recently in medical image analysis. This work proposes a deep convolutional neural network (DCNN) for Alzheimers disease classification using magnetic resonance imaging (MRI) samples. Alzheimer disease (AD) is an irreversible neurological brain disorder; its early symptoms are memory loss and losing thinking abilities called cognitive functions. The accurate diagnosis of Alzheimers disease at an early stage is very vital for patient care and conducting future treatment. Deep learning techniques are capable of learning high-level features from dataset compared to hand-crafted feature learning methods such as machine learning techniques. The proposed method classifies the disease as Alzheimers disease (AD), mild cognitive impairment (MCI) and normal control (NC). Spyder software obtained from anaconda bundle with Keras library and Tensorflow backend on GPU is used to model DCNN. Experiments are conducted using ADNI dataset and output classification result showed 98.57% accuracy compared to other studies. Our approach also enables us to expand this methodology to predict for more stages of disease classification.
Keywords:
Alzheimers disease
Convolutional neural network
Deep learning
MRI
Neurological disorder
Author(s) Name: Swathi S. Kundaram, Ketki C. Pathak
Journal name:
Conferrence name: Proceedings of the Fourth International Conference on Microelectronics, Computing and Communication Systems
Publisher name: Springer
DOI: 10.1007/978-981-15-5546-6_50
Volume Information:
