English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Parametric spatial transformation models have been successfully applied toimage registration tasks. In such models, the transformation of interest isparameterized by a fixed set of basis functions as for example B-splines. Eachbasis function is located on a fixed regular grid position among the imagedomain, because the transformation of interest is not known in advance. As aconsequence, not all basis functions will necessarily contribute to the finaltransformation which results in a non-compact representation of thetransformation. We reformulate the pairwise registration problem as a recursivesequence of successive alignments. For each element in the sequence, a localdeformation defined by its position, shape, and weight is computed by ourrecurrent registration neural network. The sum of all local deformations yieldthe final spatial alignment of both images. Formulating the registrationproblem in this way allows the network to detect non-aligned regions in theimages and to learn how to locally refine the registration properly. Incontrast to current non-sequence-based registration methods, our approachiteratively applies local spatial deformations to the images until the desiredregistration accuracy is achieved. We trained our network on 2D magneticresonance images of the lung and compared our method to a standard parametricB-spline registration. The experiments show, that our method performs on parfor the accuracy but yields a more compact representation of thetransformation. Furthermore, we achieve a speedup of around 15 compared to theB-spline registration.

Recurrent Registration Neural Networks for Deformable Image Registration