在本文中,我们提出了一种结构化表示,称为Pose as Compositional Tokens(PCT),以探索联合依赖性。它由M个离散的标记表示一个姿势,每个标记都表征一个具有几个相互依赖关节的子结构(见图1)。该组成设计使其能够以低成本实现小的重建误差。然后,我们将姿态估计作为一项分类任务。特别地,我们学习了一个分类器来预测...
Human Pose As Compositional Tokens Zigang Geng, Chunyu Wang, Yixuan Wei, Ze Liu, Houqiang Li, Han Hu June 2023 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Related File Human pose is typically represented by a coordinate vector of body joints or th...
In this paper, we present a structured representation, named Pose as Compositional Tokens (PCT), to explore the joint dependency. It represents a pose by M discrete tokens with each characterizing a sub-structure with several interdependent joints. The compositional design enables it to achieve a ...
结合我们发表在CVPR23上的文章《PCT: Human Pose as Compositional Tokens》,简单谈谈结构化表达在人体姿态估计中的作用。 论文arxiv.org/pdf/2303.1163 项目sites.google.com/view/p 代码github.com/Gengzigang/P 大部分视觉任务的输出都是结构化的,比如在human pose estimation(HPE)任务里,关键点之间就存在着很强...
Citation @inproceedings{Geng23PCT, author={Zigang Geng and Chunyu Wang and Yixuan Wei and Ze Liu and Houqiang Li and Han Hu}, title={Human Pose as Compositional Tokens}, booktitle={{CVPR}}, year={2023}, } About This is an official implementation of our CVPR 2023 paper "Human Pose as...
(2018). Deeply Learned Compositional Models for Human Pose Estimation: 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part III. Computer Vision - ECCV 2018.W. Tang, P. Yu, and Y. Wu. Deeply learned compositional models for human pose estimation. In ECCV, ...
As evidenced by Fig.6a and 6b, both the vanilla CNN-based VQ-VAE and Transformer-based VQ- VAE grapple with sub-optimal codebook usage. This is because the tokens in the codebook aren't inherently driven to develop a distinguishing motion representation. Yet, with...
Liang, Y. Wei, Compositional human pose regression, The IEEE International Conference on Computer Vision (2017) 2621-2630.X. Sun, J. Shang, S. Liang, and Y. Wei. Compositional hu- man pose regression. In International Conference on Com- puter Vision, 2017. 2, 3, 4, 7...
Mutual learningMacro-microBMMSEHuman pose estimationIn this paper, we propose to perform mutual learning inside the compositional model for human pose estimation. We conduct inference of the human pose in a compositional model which is composed of parts and subparts networks to better capture ...
Tokenpose: Learning keypoint tokens for human pose estimation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, BC, Canada, 11–17 October 2021; pp. 11313–11322. [Google Scholar] Sun, X.; Xiao, B.; Wei, F.; Liang, S.; Wei, Y. Integral human ...