Multi-class Alzheimer Disease Classification using Hybrid Features

Alzheimer disease (AD) is one of the most common form of dementia. Accurate detection of AD and its initial stage i.e., mild cognitive impairment (MCI) is a challenging task. In this study, a computer-aided diagnosis (CAD) system is implemented on clinical and diagnostic imaging data from OASIS database. Amygdala and hippocampus are the regions that are most af-fected by Alzheimer and are located inside the grey matter region of brain. Features used for classification are calculated using grey level co-occurrence matrix (GLCM) such as entropy, energy, homogeneity, and correlation. The ratios of the grey matter and white matter volume to the cerebrospinal fluid volume are also used. Clinical features are also used improving the classification accuracy achieving 94.6% for binary classification. The proposed algorithm is also used for multi-class classification where three classes, namely, normal (N), Alzheimer disease (AD), and mild cognitive impairment (MCI) are considered. An accuracy of 79.8% on these classes is achieved that is significant since the classes considered are highly similar. We have achieved improved results in comparison to state-of-the-art techniques for binary classification and have also performed multi-class classification.

@article{Anwar2018,
  title     = {Multi-class Alzheimer Disease Classification using Hybrid Features},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {9},
  number    = {1},
  year      = {2018},
  publisher = {The Science and Information Organization},
  author    = {Syed Muhammad Anwar},
  doi       = {10.14569/SpecialIssue.2018.090135},
  url       = {https://doi.org/10.14569/SpecialIssue.2018.090135}
}