Meeting Banner
Abstract #4326

Searching for new Dementia-related Features within MRI: Keypoint Detection and Description

Elisabeth Sthler 1

1 Department of Computer and Information Science, University of Konstanz, Konstanz, Baden-Wrttemberg, Germany

New dementia-related features are presented to differentiate between various stages of Alzheimers disease. Prior registration of MRI-scans, possibly unsuccessful and always time-consuming, is avoided by employing local invariant features which are independent of image scale and orientation. Feature detectors are implemented based on scale-space theory in an automatized image processing workflow, and tested on a standardized MRI collection comprising 382 T1 MRI scans from patients with Alzheimer's Disease or mild cognitive impairment, and from a control group. The approach is not only very efficient for processing large datasets, but also first order statistics of features already differentiate significantly between classes.

This abstract and the presentation materials are available to members only; a login is required.

Join Here