Alzheimer's disease diagnosis based on anatomically stratified texture analysis of the hippocampus in structural MRI | IEEE Conference Publication | IEEE Xplore

Alzheimer's disease diagnosis based on anatomically stratified texture analysis of the hippocampus in structural MRI


Abstract:

In this paper, we propose a novel method to integrate texture and shape analysis in an anatomically-informed way to model the hippocampus from structural MRI. Specificall...Show More

Abstract:

In this paper, we propose a novel method to integrate texture and shape analysis in an anatomically-informed way to model the hippocampus from structural MRI. Specifically, Poisson map is used to stratify the segmented hippocampus into layers, and a bag-of-words (BoW) model on the image intensity is generated to analyze the texture information embedded over the layers. Another level of BoW model is used to further encode the texture over subjects and across layers. A Bayesian non-parametric (BNP) method is proposed to infer and adjust for the variability of different layers. We validate our model on hippocampus-based Alzheimer's disease (AD) and mild cognitive impairment (MCI) diagnosis. A thorough comparison of related features is performed. Results demonstrate that our method achieves state-of-the-art performance while using information from the hippocampus only.
Date of Conference: 04-07 April 2018
Date Added to IEEE Xplore: 24 May 2018
ISBN Information:
Electronic ISSN: 1945-8452
Conference Location: Washington, DC, USA

References

References is not available for this document.