Variational Feature Extraction of Two-Dimensional Variational Mode Decomposition for Alzheimer’s Disease Classification

Image decomposition plays an important role in Alzheimer’s disease classification. Existing image decomposition methods can extract features but their performance is insufficiently accurate due to loss of major characteristics of brain shape in image decomposition. In this paper, a variational feature extraction of two-dimensional variational mode decomposition is proposed for classifying Alzheimer’s disease from brain MRI images. This method extracts and selects the variational features of the major characteristics of the brain, and then eliminates some features which lose the brain information. For 60 brain images from the OASIS database, the proposed method yields 94.44%, higher accuracy than the state-of-the-art methods. Our method is able to extract the variational features with the major characteristics of the brain, thereby improving the performance of Alzheimer’s disease classification.