Saved in:
Bibliographic Details
Main Authors: Le, Cuong, Melnyk, Pavlo, Wandt, Bastian, Wadenbäck, Mårten
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.16763
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910251705434112
author Le, Cuong
Melnyk, Pavlo
Wandt, Bastian
Wadenbäck, Mårten
author_facet Le, Cuong
Melnyk, Pavlo
Wandt, Bastian
Wadenbäck, Mårten
contents Recovering 3D human poses from a monocular camera view is a highly ill-posed problem due to the depth ambiguity. Earlier studies on 3D human pose lifting from 2D often contain incorrect-yet-overconfident 3D estimations. To mitigate the problem, emerging probabilistic approaches treat the 3D estimations as a distribution, taking into account the uncertainty measurement of the poses. Falling in a similar category, we proposed FMPose, a probabilistic 3D human pose estimation method based on the flow matching generative approach. Conditioned on the 2D cues, the flow matching scheme learns the optimal transport from a simple source distribution to the plausible 3D human pose distribution via continuous normalizing flows. The 2D lifting condition is modeled via graph convolutional networks, leveraging the learnable connections between human body joints as the graph structure for feature aggregation. While trade-offs between processing time and precision exist, already in the equal-accuracy comparison, FMPose exhibits significantly faster processing time than the diffusion model, and also offers another faster and more accurate configuration. Experimental results show major improvements of our FMPose over current state-of-the-art methods on two common benchmarks for 3D human pose estimation, namely Human3.6M, MPI-INF-3DHP. Additionally, FMPose shows competitive performance on the more challenging 3DPW dataset. The code implementation is available at https://github.com/cuongle1206/FMPose
format Preprint
id arxiv_https___arxiv_org_abs_2601_16763
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Flow Matching for Probabilistic Monocular 3D Human Pose Estimation
Le, Cuong
Melnyk, Pavlo
Wandt, Bastian
Wadenbäck, Mårten
Computer Vision and Pattern Recognition
Recovering 3D human poses from a monocular camera view is a highly ill-posed problem due to the depth ambiguity. Earlier studies on 3D human pose lifting from 2D often contain incorrect-yet-overconfident 3D estimations. To mitigate the problem, emerging probabilistic approaches treat the 3D estimations as a distribution, taking into account the uncertainty measurement of the poses. Falling in a similar category, we proposed FMPose, a probabilistic 3D human pose estimation method based on the flow matching generative approach. Conditioned on the 2D cues, the flow matching scheme learns the optimal transport from a simple source distribution to the plausible 3D human pose distribution via continuous normalizing flows. The 2D lifting condition is modeled via graph convolutional networks, leveraging the learnable connections between human body joints as the graph structure for feature aggregation. While trade-offs between processing time and precision exist, already in the equal-accuracy comparison, FMPose exhibits significantly faster processing time than the diffusion model, and also offers another faster and more accurate configuration. Experimental results show major improvements of our FMPose over current state-of-the-art methods on two common benchmarks for 3D human pose estimation, namely Human3.6M, MPI-INF-3DHP. Additionally, FMPose shows competitive performance on the more challenging 3DPW dataset. The code implementation is available at https://github.com/cuongle1206/FMPose
title Flow Matching for Probabilistic Monocular 3D Human Pose Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.16763