FMPose3D：基于流匹配的单目三维姿态估计
FMPose3D: monocular 3D pose estimation via flow matching
把流匹配的少步 ODE 采样引入单目 3D 姿态估计，用不同噪声种子生成多假设、再用重投影后验期望聚合出单一结果，可直接迁移到任何存在深度歧义的 2D→3D 生成式回归任务（手部、动物、物体 6D 位姿）。
原文依据 we leverage Flow Matching (FM) to learn a velocity field defined by an Ordinary Differential Equation (ODE)
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论由于深度歧义与遮挡，单目三维姿态估计本质上是病态问题，因而促使人们采用能够生成多个合理三维姿态假设的概率方法。特别是，基于扩散的模型近来表现出强劲性能，但其迭代去噪过程对每次预测通常需要大量时间步，导致推理计算代价高昂。相比之下，我们利用流匹配（FM）学习由常微分方程（ODE）定义的速度场，从而仅需少量积分步即可高效生成三维姿态样本。我们提出一种新颖的生成式姿态估计框架 FMPose3D，将三维姿态估计表述为条件分布传输问题。它持续地将样本从标准高斯先验传输到仅以二维输入为条件的合理三维姿态分布。尽管 ODE 轨迹是确定性的，FMPose3D 通过采样不同的噪声种子自然地生成多样的姿态假设。为了从这些假设中得到单一准确预测，我们进一步引入基于重投影的后验期望聚合（RPEA）模块，它近似三维假设上的贝叶斯后验期望。FMPose3D 在广泛使用的人体姿态估计基准 Human3.6M 与 MPI-INF-3DHP 上超越现有方法，并在三维动物姿态数据集 Animal3D 与 CtrlAni3D 上进一步取得最优性能，在两类三维姿态领域均展现出强劲表现。代码见 https://github.com/AdaptiveMotorControlLab/FMPose3D。
英文原文摘要 1627 字符
Monocular 3D pose estimation is fundamentally ill-posed due to depth ambiguity and occlusions, thereby motivating probabilistic methods that generate multiple plausible 3D pose hypotheses. In particular, diffusion-based models have recently demonstrated strong performance, but their iterative denoising process typically requires many timesteps for each prediction, making inference computationally expensive. In contrast, we leverage Flow Matching (FM) to learn a velocity field defined by an Ordinary Differential Equation (ODE), enabling efficient generation of 3D pose samples with only a few integration steps. We propose a novel generative pose estimation framework, FMPose3D, that formulates 3D pose estimation as a conditional distribution transport problem. It continuously transports samples from a standard Gaussian prior to the distribution of plausible 3D poses conditioned only on 2D inputs. Although ODE trajectories are deterministic, FMPose3D naturally generates various pose hypotheses by sampling different noise seeds. To obtain a single accurate prediction from those hypotheses, we further introduce a Reprojection-based Posterior Expectation Aggregation (RPEA) module, which approximates the Bayesian posterior expectation over 3D hypotheses. FMPose3D surpasses existing methods on the widely used human pose estimation benchmarks Human3.6M and MPI-INF-3DHP, and further achieves state-of-the-art performance on the 3D animal pose datasets Animal3D and CtrlAni3D, demonstrating strong performance across both 3D pose domains. The code is available at https://github.com/AdaptiveMotorControlLab/FMPose3D.