GeoPose framework enables rapid, patient-agnostic CTA-to-DSA registration

Researchers have developed GeoPose, a novel framework for aligning intraoperative digital subtraction angiography (DSA) with pre-procedural computed tomography angiography (CTA). Unlike existing methods that are sensitive to initialization or require patient-specific training, GeoPose utilizes a population-trained residual network to estimate C-arm pose in a canonical frame and then transfers it to the CTA's native frame. This approach achieves rapid and accurate registration, with an optimization-free version reaching a mean projected centerline distance of 5.8 mm in 0.15 seconds, significantly outperforming baseline methods. AI

IMPACT This new framework could accelerate and improve the accuracy of image registration in medical procedures, potentially leading to better surgical outcomes.

RANK_REASON This is a research paper detailing a new technical framework for medical imaging registration. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.CV →

• arXiv
• computed tomography angiography
• digital subtraction angiography
• GeoPose
• Rudolf L.M. Van Herten

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →