eduzhai > Applied Sciences > Engineering >

Overview of Scanner Invariant Representations

  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: Pooled imaging data from multiple sources is subject to bias from eachsource. Studies that do not correct for these scanner site biases at best losestatistical power, and at worst leave spurious correlations in their data.Estimation of the bias effects is non-trivial due to the paucity of data withcorrespondence across sites, so called "traveling phantom " data, which isexpensive to collect. Nevertheless, numerous solutions leveraging directcorrespondence have been proposed. In contrast to this, Moyer et al. (2019)proposes an unsupervised solution using invariant representations, one whichdoes not require correspondence and thus does not require paired images. Byleveraging the data processing inequality, an invariant representation can thenbe used to create an image reconstruction that is uninformative of its originalsource, yet still faithful to the underlying structure. In the present abstractwe provide an overview of this method.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...