Human Mesh Recovery from Arbitrary Multi-view Images

📄 arXiv: 2403.12434v4 📥 PDF

作者: Xiaoben Li, Mancheng Meng, Ziyan Wu, Terrence Chen, Fan Yang, Dinggang Shen

分类: cs.CV

发布日期: 2024-03-19 (更新: 2024-06-17)


💡 一句话要点

提出统一人类网格恢复框架以解决多视角图像问题

🎯 匹配领域: 支柱六:视频提取与匹配 (Video Extraction)

关键词: 人类网格恢复 多视角图像 相机姿态估计 深度学习 变换解码器 计算机视觉 虚拟现实

📋 核心要点

  1. 现有方法在同时估计任意相机姿态和恢复人类网格时面临灵活性不足的挑战。
  2. 本文提出的U-HMR框架通过解耦相机姿态和人类网格的估计,分别采用两个子网络处理。
  3. 在多个公共数据集上进行的实验表明,U-HMR在性能上显著优于现有方法,展示了其有效性。

📝 摘要(中文)

从任意多视角图像中恢复人类网格涉及两个特征:任意相机姿态和任意数量的相机视角。由于这些变异性,设计一个统一框架来解决这一任务具有挑战性。为此,本文提出了一种分而治之的框架——统一人类网格恢复(U-HMR),该框架由相机与身体解耦、相机姿态估计和任意视角融合三个主要组件组成。通过将相机姿态和人类网格的估计分为两个子任务,U-HMR实现了灵活的处理。实验结果表明,该框架在多个公共数据集上表现出色,验证了其有效性和灵活性。

🔬 方法详解

问题定义:本文旨在解决从任意多视角图像中恢复人类网格的问题。现有方法在处理任意相机姿态和多视角信息时,往往缺乏灵活性,难以同时满足这两个要求。

核心思路:论文提出的U-HMR框架通过将相机姿态估计与人类网格恢复解耦,分别由两个子网络处理,从而提高了整体的灵活性和准确性。

技术框架:U-HMR框架主要包括三个模块:相机与身体解耦(CBD)、相机姿态估计(CPE)和任意视角融合(AVF)。CBD将相机姿态和网格恢复分为两个独立的子任务,CPE并行处理所有视角的相机姿态,而AVF则通过变换解码器融合多视角信息。

关键创新:最重要的创新在于将相机姿态和人类网格的估计解耦,利用共享的多层感知机(MLP)并行处理相机姿态,同时引入变换解码器以提取跨视角特征,显著提升了恢复效果。

关键设计:在CPE中,采用共享MLP处理所有视角的相机姿态;在AVF中,使用带有SMPL参数查询标记的变换解码器,以实现视角信息的有效融合,确保融合操作与视角数量无关。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

在Human3.6M、MPI-INF-3DHP和TotalCapture等三个公共数据集上的实验结果显示,U-HMR框架在网格恢复精度上相较于现有方法有显著提升,具体性能数据未提供,但实验表明其有效性和灵活性。

🎯 应用场景

该研究的潜在应用领域包括虚拟现实、增强现实、运动分析和人机交互等。通过准确恢复人类网格,能够提升这些领域中的用户体验和交互效果,具有重要的实际价值和广泛的未来影响。

📄 摘要(原文)

Human mesh recovery from arbitrary multi-view images involves two characteristics: the arbitrary camera poses and arbitrary number of camera views. Because of the variability, designing a unified framework to tackle this task is challenging. The challenges can be summarized as the dilemma of being able to simultaneously estimate arbitrary camera poses and recover human mesh from arbitrary multi-view images while maintaining flexibility. To solve this dilemma, we propose a divide and conquer framework for Unified Human Mesh Recovery (U-HMR) from arbitrary multi-view images. In particular, U-HMR consists of a decoupled structure and two main components: camera and body decoupling (CBD), camera pose estimation (CPE), and arbitrary view fusion (AVF). As camera poses and human body mesh are independent of each other, CBD splits the estimation of them into two sub-tasks for two individual sub-networks (ie, CPE and AVF) to handle respectively, thus the two sub-tasks are disentangled. In CPE, since each camera pose is unrelated to the others, we adopt a shared MLP to process all views in a parallel way. In AVF, in order to fuse multi-view information and make the fusion operation independent of the number of views, we introduce a transformer decoder with a SMPL parameters query token to extract cross-view features for mesh recovery. To demonstrate the efficacy and flexibility of the proposed framework and effect of each component, we conduct extensive experiments on three public datasets: Human3.6M, MPI-INF-3DHP, and TotalCapture.