Tensor optimal scoring for alzheimer's disease detection

In recent years, the universal application of magnetic resonance imaging has given a new diagnostic approach for lots of diseases, including Alzheimer's disease (AD). In this paper, we propose a new classification framework for AD detection. To avoid the curse of dimensionality problem, a new dimension reduction algorithm in tensor space called tensor optimal scoring (TOS) was proposed to extract efficient features from MRI data of AD. Alternating least square optimization was employed to estimate the projection vector and optimal scores. We validated the effectiveness of our proposed method on OASIS dataset. The experimental results showed that our method outperformed the traditional methods especially for higher order MRI data.