Text-driven 3D human pose estimation

Paper Title:

FinePOSE: Fine-Grained Prompt-Driven 3D Human Pose Estimation via Diffusion Models

Published on:

8 May 2024

Primary Category:

Computer Vision and Pattern Recognition

Paper Authors:

Jinglin Xu,

Yijie Guo,

Yuxin Peng

Bullets

Key Details

•

Introduces a novel fine-grained part-aware prompt learning mechanism

•

Establishes fine-grained communications between prompts and poses

•

Integrates prompt embeddings and noise levels to enable adaptive denoising

•

Achieves state-of-the-art on Human3.6M and MPI-INF-3DHP datasets

•

Shows potential for complex multi-human pose estimation

Explore the topics in this paper

3d human pose estimation

diffusion models

keypoints

multi-person pose estimation

prompts

AI generated summary

Text-driven 3D human pose estimation

This paper proposes FinePOSE, a new diffusion model-based approach for estimating 3D human poses from 2D keypoints. It introduces a novel fine-grained part-aware prompt learning mechanism to provide precise guidance for each human body part's movement. FinePOSE also establishes communications between the learned prompts and poses to enhance the diffusion model's denoising capability. Experiments show state-of-the-art performance on public benchmarks. An extension to multi-human scenarios also demonstrates potential.