E(3)-Pose: physics-aware deep learning to track fetal head motion in MRI
Researchers at MIT, Boston Children’s Hospital, and Inria developed E(3)-Pose, a deep learning model that embeds physical priors (notably left–right symmetry/equivariance) into its architecture to track fetal head motion during MRI and help assemble higher-resolution volumetric scans; the work was presented at ECCV 2026. The approach aims to reduce coverage gaps and orientation errors in fetal MRI that can lead to missed or incorrect prenatal brain diagnoses.