Camera Distance-Aware Top-Down Approach for 3D Multi-Person Pose Estimation From a Single RGB Image

By Gyeongsik Moon, Ju Yong Chang, Kyoung Mu Lee. IEEE International Conference on Computer Vision, 2019

376 citations recorded by the source corpus.

Although significant improvement has been achieved recently in 3D human pose estimation, most of the previous methods only treat a single-person case. In this work, we firstly propose a fully learning-based, camera distance-aware top-down approach for 3D multi-person pose estimation from a single RGB image. The pipeline of the proposed system consists of human detection, absolute 3D human root localization, and root-relative 3D single-person pose estimation modules. Our system achieves comparable results with the state-of-the-art 3D single-person pose estimation models without any groundtruth information and significantly outperforms previous 3D multi-person pose estimation methods on publicly available datasets. The code is available in \footnote{\url{https://github.com/mks0601/3DMPPE_ROOTNET_RELEASE}}\textsuperscript{,}\footnote{\url{https://github.com/mks0601/3DMPPE_POSENET_RELEASE}}.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/213e0b84c93d0ac8d96a257b11e3a4891bae4a40

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose (2020, background): “We also fed state-of-the-art 3D pose outputs from [42] to (d), to validate the practical potential for performance improvement.”
Two-hand Global 3D Pose Estimation using Monocular RGB (2020, methodology): “Unlike other methods [11,16] that attempt to estimate the approximate global poses, our algorithm is, to our knowledge, the first capable of perfectly reconstructing the global 3D poses.”
Deep 3D human pose estimation: A review (2021, positive): “For multi-person estimation from a single monocular image, we introduce three two-stage methods (Rogez et al., 2019; Moon et al., 2019; Mehta et al., 2018) and a new single-stage method (Nie et al., 2019).”
I2L-MeshNet: Image-to-Lixel Prediction Network for Accurate 3D Human Pose and Mesh Estimation from a Single RGB Image (2020, methodology): “MPJPE and PA MPJPE are used for the evaluation [33], which is Euclidean distance (mm) between predicted and groundtruth 3D joint coordinates after root joint alignment and further rigid alignment, respectively.”
Human Pose Regression with Residual Log-likelihood Estimation (2021, methodology): “Method Sun [53] Nibali [41] Sun [54] Moon [35] Zhou [69] Ours”
Boosting Image-based Mutual Gaze Detection using Pseudo 3D Gaze (2020, background): “This estimate roughly corresponds to 53 ◦ horizontal Field of View (FoV): Motivated by (Moon, Chang, and Lee 2019), we assume that the depth of a head is inversely proportional to the square root of the corresponding 2D bounding box area: where α is a proportionality constant and A i denotes the 2D…”
MotionBERT: A Unified Perspective on Learning Human Motion Representations (2022, background): “The first is to estimate 3D poses with CNN directly from images [82, 104, 136].”
Deep Learning-based Human Pose Estimation: A Survey (2020, methodology): “[166] introduced a camera distance-aware approach under top-down pipeline.”