Flow Matching for Probabilistic Monocular 3D Human Pose Estimation

2026-01-23

AI Summary

This work addresses the challenge of depth ambiguity in monocular 3D human pose estimation, which often leads to overconfident yet erroneous predictions. To this end, we propose FMPose, a probabilistic 3D pose estimation framework based on flow matching that, for the first time, integrates optimal transport and continuous normalizing flows into this task. FMPose employs a graph convolutional network to model the lifting condition from 2D to 3D poses and learns an optimal transport path from a simple source distribution to a plausible 3D pose distribution. Compared to diffusion-based approaches, FMPose achieves superior accuracy and computational efficiency. Extensive experiments demonstrate that FMPose significantly outperforms state-of-the-art methods on three major benchmarks: Human3.6M, MPI-INF-3DHP, and 3DPW.

Technology Category

• Computer Vision: 3D Computer Vision
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: State Estimation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

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. Compared to diffusion-based methods, the FMPose with optimal transport produces faster and more accurate 3D pose generations. Experimental results show major improvements of our FMPose over current state-of-the-art methods on three common benchmarks for 3D human pose estimation, namely Human3.6M, MPI-INF-3DHP and 3DPW.

Problem

• 3D human pose estimation
• monocular vision
• depth ambiguity
• probabilistic modeling
• ill-posed problem

Innovation

• flow matching
• probabilistic 3D pose estimation
• optimal transport
• graph convolutional networks
• monocular 3D human pose