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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Motion estimation is a fundamental step in dynamic medical image processingfor the assessment of target organ anatomy and function. However, existingimage-based motion estimation methods, which optimize the motion field byevaluating the local image similarity, are prone to produce implausibleestimation, especially in the presence of large motion. In this study, weprovide a novel motion estimation framework of Dense-Sparse-Dense (DSD), whichcomprises two stages. In the first stage, we process the raw dense image toextract sparse landmarks to represent the target organ anatomical topology anddiscard the redundant information that is unnecessary for motion estimation.For this purpose, we introduce an unsupervised 3D landmark detection network toextract spatially sparse but representative landmarks for the target organmotion estimation. In the second stage, we derive the sparse motiondisplacement from the extracted sparse landmarks of two images of differenttime points. Then, we present a motion reconstruction network to construct themotion field by projecting the sparse landmarks displacement back into thedense image domain. Furthermore, we employ the estimated motion field from ourtwo-stage DSD framework as initialization and boost the motion estimationquality in light-weight yet effective iterative optimization. We evaluate ourmethod on two dynamic medical imaging tasks to model cardiac motion and lungrespiratory motion, respectively. Our method has produced superior motionestimation accuracy compared to existing comparative methods. Besides, theextensive experimental results demonstrate that our solution can extract wellrepresentative anatomical landmarks without any requirement of manualannotation. Our code is publicly available online.

Unsupervised Landmark Detection Based Spatiotemporal Motion Estimation for 4D Dynamic Medical Images