首页|基于深度学习的二维人体姿态估计研究进展

基于深度学习的二维人体姿态估计研究进展

扫码查看
人体姿态估计在人体行为识别、人机交互、体育运动分析等方面有着广泛的应用前景,是计算机视觉领域的一个研究热点.在最近的十年中,得益于深度学习技术,大量的研究工作极大地推动了人体姿态估计技术的发展,但由于受训练样本不足、人体姿态的多变性、遮挡、环境的复杂性等因素影响,人体姿态估计仍然面临着诸多的挑战.文中对近年来基于深度学习的2D人体姿态估计方法进行归纳和总结,着重分析一些有代表性的人体姿态估计方法的思路及工作原理,以便研究人员了解当前的研究现状、面临的挑战以及今后的研究方向,拓展研究思路.
Research progress on two-dimensional human pose estimation based on deep learning
Human pose estimation has broad application prospects in human behavior recognition,human computer interaction,sports analysis,and has been a research hotspot in the field of computer vision.In the past decade,thanks to deep learning technology,a large amount of studies have greatly promoted the development of human pose estimation technology.However,due to factors such as insufficient training samples,variability of human pose,occlusion,and complexity of the environment,human pose estimation still faces many challenges.This paper reviews and summarizes the 2D human pose estimation methods based on deep learning in recent years,focusing on analyzing the ideas and principles of some representative human pose estimation methods,so that scholars in this field can understand current research status,challenges,and future directions,and expand their research ideas.

human pose estimation(HPE)single-person pose estimationmulti-person pose estimationdeep learningkeypoint detection

卢官明、卢峻禾、陈晨

展开 >

南京邮电大学通信与信息工程学院,江苏南京 210003

南京邮电大学计算机学院,江苏南京 210023

人体姿态估计 单人体姿态估计 多人体姿态估计 深度学习 关键点检测

国家自然科学基金

72074038

2024

南京邮电大学学报(自然科学版)
南京邮电大学

南京邮电大学学报(自然科学版)

CSTPCD北大核心
影响因子:0.486
ISSN:1673-5439
年,卷(期):2024.44(1)
  • 72