Diagnosis of Alzheimer's disease from MR images using relevance feedback

Abstract

In this work, we present a learning framework to help early diagnosis of Alzheimer's disease (AD) from magnetic resonance images. Our approach relies on a nearest neighbor (NN) procedure where the similarity measure is obtained via on-line supervised learning. We propose two alternative approaches to learn the similarities between cases. Several experiments on OASIS database establish that, even with weak global visual descriptors and small training sets, this framework has better diagnostic performance than standard classification based approaches and enjoys a certain degree of robustness against incorrect relevance judgments.