← Back to all news
arXiv Computer Vision By R. James Cotton, Divya Joshi, Colleen Peyton

Cross-Model Distillation of a Human-Pose Foundation Model from Unannotated Infant Video for Markerless 3D Pose Estimation

Read the original on arXiv Computer Vision →

The paper presents a method for improving markerless 3D pose estimation in infants by cross‑model distillation. Using unannotated infant video, a frozen Sapiens 2 pose model provides dense pseudo‑labels that guide fine‑tuning of the SAM 3D Body model. On a held‑out dataset of eleven infants, the fine‑tuned model shows significant gains in 2D keypoint accuracy and 3D joint error compared to the original SAM 3D Body model.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.

Related stories

arXiv Computer Vision
Sep 17

Video-Based Markerless Motion Capture for Clinical and Rehabilitation Biomechanics: A PRISMA-ScR Scoping Review of Validated Architectures, Clinical Readiness, and Emerging Methods

This scoping review examined 117 studies on video-based markerless motion capture, most published from 2024 onward and focused on healthy adults walking in laboratories. The studies identified five main pipeline architectures, but most reported only raw joint angles without biomechanical refinement, achieving sagittal lower‑limb agreement of about 5–6°, which falls short of clinical acceptability. Validation of out‑of‑plane kinematics, kinetics, and performance in older or pathological populations was rare, and emerging computer‑vision techniques such as foundation‑model mesh recovery and differentiable inverse kinematics were largely absent from validated work.

By Florian Delaplace (LAMHESS, CHU), Elodie Piche (LAMHESS), Fr\'ed\'eric Chorin (IUF, LAMHESS), Raphael Zory (IUF, LAMHESS)
More like this →