Pose Prior

Pose priors, representing prior knowledge about the likely configurations of poses (e.g., human or animal postures), are crucial for improving the accuracy and robustness of pose estimation and generation models. Current research focuses on learning these priors in unsupervised or self-supervised ways from video or image data, employing various architectures such as neural distance fields, graph neural networks, and diffusion models to capture the complex relationships between pose components. These advancements enable more accurate pose estimation even in challenging scenarios like occlusion or limited data, with applications ranging from human-computer interaction to animation and robotics. The development of effective pose priors is significantly impacting the field by enabling more realistic and accurate 3D pose estimation and generation.

Papers

October 18, 2024

Multi-modal Pose Diffuser: A Multimodal Generative Conditional Pose Prior
Calvin-Khang Ta, Arindam Dutta, Rohit Kundu, Rohit Lal, Hannah Dela Cruz, Dripta S. Raychaudhuri, Amit Roy-Chowdhury
Conditional Generative Multi Person Modal Diffusion Pose Prior Pose Diversity

October 4, 2024

Unsupervised Prior Learning: Discovering Categorical Pose Priors from Videos
Ziyu Wang, Shuangpeng Han, Mike Zheng Shou, Mengmi Zhang
Pose Estimation Gameplay Video Unsupervised Pre Training Category Level Pose Prior

August 29, 2024

GRPose: Learning Graph Relations for Human Image Generation with Pose Priors
Xiangchen Yin, Donglin Di, Lei Fan, Hao Li, Chen Wei, Xiaofei Gou, Yang Song, Xiao Sun, Xun Yang
Relational Graph Human Image Generation Pose Prior

May 6, 2024

Pose Priors from Language Models
Sanjay Subramanian, Evonne Ng, Lea Müller, Dan Klein, Shiry Ginosar, Trevor Darrell
Language Model Large Multimodal Model Human Annotation 3D Pose Physical Interaction Contact State Estimation Pose Prior

March 5, 2024

NRDF: Neural Riemannian Distance Fields for Learning Articulated Pose Priors
Yannan He, Garvita Tiwari, Tolga Birdal, Jan Eric Lenssen, Gerard Pons-Moll
Distance Field Riemannian Gradient Pose Prior Riemannian Flow Matching

November 7, 2023

UP-NeRF: Unconstrained Pose-Prior-Free Neural Radiance Fields
Injae Kim, Minhyuk Choi, Hyunwoo J. Kim
Neural Radiance Field Novel View Synthesis Depth Supervision Unconstrained Image Pose Prior

June 8, 2023

LU-NeRF: Scene and Pose Estimation by Synchronizing Local Unposed NeRFs
Zezhou Cheng, Carlos Esteves, Varun Jampani, Abhishek Kar, Subhransu Maji, Ameesh Makadia
Neural Radiance Field Pose Estimation NeRF SLAM 3D Human Pose Scene Structure Synchronization Parameter Update Barrier Based NeRF Unusual Pose Pose Prior

January 18, 2023

Learning 3D-aware Image Synthesis with Unknown Pose Distribution
Zifan Shi, Yujun Shen, Yinghao Xu, Sida Peng, Yiyi Liao, Sheng Guo, Qifeng Chen, Dit-Yan Yeung
LeArning Abstract Image Synthesis Pose Generation Pose Distribution Generative Radiance Field Pose Prior

July 29, 2022

Explicit Occlusion Reasoning for Multi-person 3D Human Pose Estimation
Qihao Liu, Yi Zhang, Song Bai, Alan Yuille
Human Pose Estimation Multi Person Multi Person 3D Pose Estimation Pose Prior Occlusion Reasoning

July 27, 2022

Pose-NDF: Modeling Human Pose Manifolds with Neural Distance Fields
Garvita Tiwari, Dimitrije Antic, Jan Eric Lenssen, Nikolaos Sarafianos, Tony Tung, Gerard Pons-Moll
Distance Field Plausible Pose Pose Prior