ADiag: Graph Neural Network Based Diagnosis of Alzheimer's Disease
Authors:
Vishnu Ram Sampathkumar
Abstract:
Alzheimer's Disease (AD) is the most widespread neurodegenerative disease, affecting over 50 million people across the world. While its progression cannot be stopped, early and accurate diagnostic testing can drastically improve quality of life in patients. Currently, only qualitative means of testing are employed in the form of scoring performance on a battery of cognitive tests. The inherent dis…
▽ More
Alzheimer's Disease (AD) is the most widespread neurodegenerative disease, affecting over 50 million people across the world. While its progression cannot be stopped, early and accurate diagnostic testing can drastically improve quality of life in patients. Currently, only qualitative means of testing are employed in the form of scoring performance on a battery of cognitive tests. The inherent disadvantage of this method is that the burden of an accurate diagnosis falls on the clinician's competence. Quantitative methods like MRI scan assessment are inaccurate at best,due to the elusive nature of visually observable changes in the brain. In lieu of these disadvantages to extant methods of AD diagnosis, we have developed ADiag, a novel quantitative method to diagnose AD through GraphSAGE Network and Dense Differentiable Pooling (DDP) analysis of large graphs based on thickness difference between different structural regions of the cortex. Preliminary tests of ADiag have revealed a robust accuracy of 83%, vastly outperforming other qualitative and quantitative diagnostic techniques.
△ Less
Submitted 8 January, 2021;
originally announced January 2021.
Distributed optimization for nonrigid nano-tomography
Authors:
Viktor Nikitin,
Vincent De Andrade,
Azat Slyamov,
Benjamin J. Gould,
Yuepeng Zhang,
Vandana Sampathkumar,
Narayanan Kasthuri,
Doga Gursoy,
Francesco De Carlo
Abstract:
Resolution level and reconstruction quality in nano-computed tomography (nano-CT) are in part limited by the stability of microscopes, because the magnitude of mechanical vibrations during scanning becomes comparable to the imaging resolution, and the ability of the samples to resist beam damage during data acquisition. In such cases, there is no incentive in recovering the sample state at differe…
▽ More
Resolution level and reconstruction quality in nano-computed tomography (nano-CT) are in part limited by the stability of microscopes, because the magnitude of mechanical vibrations during scanning becomes comparable to the imaging resolution, and the ability of the samples to resist beam damage during data acquisition. In such cases, there is no incentive in recovering the sample state at different time steps like in time-resolved reconstruction methods, but instead the goal is to retrieve a single reconstruction at the highest possible spatial resolution and without any imaging artifacts. Here we propose a joint solver for imaging samples at the nanoscale with projection alignment, unwarping and regularization. Projection data consistency is regulated by dense optical flow estimated by Farneback's algorithm, leading to sharp sample reconstructions with less artifacts. Synthetic data tests show robustness of the method to Poisson and low-frequency background noise. Applicability of the method is demonstrated on two large-scale nano-imaging experimental data sets.
△ Less
Submitted 28 February, 2021; v1 submitted 11 July, 2020;
originally announced August 2020.
