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

Many tasks performed in image-guided, mini-invasive, medical procedures canbe cast as pose estimation problems, where an X-ray projection is utilized toreach a target in 3D space. Expanding on recent advances in the differentiablerendering of optically reflective materials, we introduce new methods for poseestimation of radiolucent objects using X-ray projections, and we demonstratethe critical role of optimal view synthesis in performing this task. We firstdevelop an algorithm (DiffDRR) that efficiently computes DigitallyReconstructed Radiographs (DRRs) and leverages automatic differentiation withinTensorFlow. Pose estimation is performed by iterative gradient descent using aloss function that quantifies the similarity of the DRR synthesized from arandomly initialized pose and the true fluoroscopic image at the target pose.We propose two novel methods for high-fidelity view synthesis, Neural TunedTomography (NeTT) and masked Neural Radiance Fields (mNeRF). Both methods relyon classic Cone-Beam Computerized Tomography (CBCT); NeTT directly optimizesthe CBCT densities, while the non-zero values of mNeRF are constrained by a 3Dmask of the anatomic region segmented from CBCT. We demonstrate that both NeTTand mNeRF distinctly improve pose estimation within our framework. By defininga successful pose estimate to be a 3D angle error of less than 3 deg, we findthat NeTT and mNeRF can achieve similar results, both with overall successrates more than 93%. However, the computational cost of NeTT is significantlylower than mNeRF in both training and pose estimation. Furthermore, we showthat a NeTT trained for a single subject can generalize to synthesizehigh-fidelity DRRs and ensure robust pose estimations for all other subjects.Therefore, we suggest that NeTT is an attractive option for robust poseestimation using fluoroscopic projections.

Robust Single-view Cone-beam X-ray Pose Estimation with Neural Tuned Tomography (NeTT) and Masked Neural Radiance Fields (mNeRF)