Landmark Detection Uncertainty as a Reliability Weight for Robust Landmark-based 2D/3D Pelvic Pose Estimation

Yehyun Suh, Brayden Schott, Chou Mo, J. Ryan Martin, Daniel Moyer
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:1722-1739, 2026.

Abstract

Landmark-based 2D/3D pelvis registration is vulnerable to noisy or ambiguous landmark detections in fluoroscopy, which can destabilize downstream pose estimation. We present an uncertainty-aware registration framework that models epistemic uncertainty in predicted landmarks and incorporates it directly into the Perspective-n-Point formulation. Using Monte Carlo dropout within a U-Net detector, we compute sample-specific per-landmark reliability estimates using the variance of multiple stochastic forward passes. These reliability estimates guide two complementary strategies: continuous weighting, which integrates uncertainty into a weighted PnP optimization, and discrete selection, which removes the most uncertain landmarks during inference. We evaluate the framework on both CT-derived synthetic fluoroscopy and real fluoroscopy from DeepFluoro. Our experiments show that uncertainty provides a principled mechanism for identifying unreliable landmarks and stabilizing pose estimation, enabling more robust registration and establishing a foundation for uncertainty-guided image-guided surgical workflows.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-suh26a, title = {Landmark Detection Uncertainty as a Reliability Weight for Robust Landmark-based 2D/3D Pelvic Pose Estimation}, author = {Suh, Yehyun and Schott, Brayden and Mo, Chou and Martin, J. Ryan and Moyer, Daniel}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {1722--1739}, year = {2026}, editor = {Huo, Yuankai and Gao, Mingchen and Kuo, Chang-Fu and Jin, Yueming and Deng, Ruining}, volume = {315}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v315/main/assets/suh26a/suh26a.pdf}, url = {https://proceedings.mlr.press/v315/suh26a.html}, abstract = {Landmark-based 2D/3D pelvis registration is vulnerable to noisy or ambiguous landmark detections in fluoroscopy, which can destabilize downstream pose estimation. We present an uncertainty-aware registration framework that models epistemic uncertainty in predicted landmarks and incorporates it directly into the Perspective-n-Point formulation. Using Monte Carlo dropout within a U-Net detector, we compute sample-specific per-landmark reliability estimates using the variance of multiple stochastic forward passes. These reliability estimates guide two complementary strategies: continuous weighting, which integrates uncertainty into a weighted PnP optimization, and discrete selection, which removes the most uncertain landmarks during inference. We evaluate the framework on both CT-derived synthetic fluoroscopy and real fluoroscopy from DeepFluoro. Our experiments show that uncertainty provides a principled mechanism for identifying unreliable landmarks and stabilizing pose estimation, enabling more robust registration and establishing a foundation for uncertainty-guided image-guided surgical workflows.} }
Endnote
%0 Conference Paper %T Landmark Detection Uncertainty as a Reliability Weight for Robust Landmark-based 2D/3D Pelvic Pose Estimation %A Yehyun Suh %A Brayden Schott %A Chou Mo %A J. Ryan Martin %A Daniel Moyer %B Proceedings of The 9th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Yuankai Huo %E Mingchen Gao %E Chang-Fu Kuo %E Yueming Jin %E Ruining Deng %F pmlr-v315-suh26a %I PMLR %P 1722--1739 %U https://proceedings.mlr.press/v315/suh26a.html %V 315 %X Landmark-based 2D/3D pelvis registration is vulnerable to noisy or ambiguous landmark detections in fluoroscopy, which can destabilize downstream pose estimation. We present an uncertainty-aware registration framework that models epistemic uncertainty in predicted landmarks and incorporates it directly into the Perspective-n-Point formulation. Using Monte Carlo dropout within a U-Net detector, we compute sample-specific per-landmark reliability estimates using the variance of multiple stochastic forward passes. These reliability estimates guide two complementary strategies: continuous weighting, which integrates uncertainty into a weighted PnP optimization, and discrete selection, which removes the most uncertain landmarks during inference. We evaluate the framework on both CT-derived synthetic fluoroscopy and real fluoroscopy from DeepFluoro. Our experiments show that uncertainty provides a principled mechanism for identifying unreliable landmarks and stabilizing pose estimation, enabling more robust registration and establishing a foundation for uncertainty-guided image-guided surgical workflows.
APA
Suh, Y., Schott, B., Mo, C., Martin, J.R. & Moyer, D.. (2026). Landmark Detection Uncertainty as a Reliability Weight for Robust Landmark-based 2D/3D Pelvic Pose Estimation. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:1722-1739 Available from https://proceedings.mlr.press/v315/suh26a.html.

Related Material