摘要
本文针对基于深度学习的二维人体姿态估计方法进行了全面综述。首先对这些深度学习技术进行了分类、分析和比较,并介绍了二维人体姿态估计中常用的数据集和指标最后还讨论了有待解决的问题和未来研究的挑战。
This paper provides a comprehensive overview of the two-dimensional human pose estimation methods based on deep learning.First,these deep learning technologies are classified,analyzed and compared,and the commonly used data sets and indicators in two-dimensional human pose estimation are introduced.Finally,the problems to be solved and the challenges of future research are discussed.
作者
张静静
宁媛
章成学
ZHANG Jingjing;NING Yuan;ZHANG Chengxue(College of Electrical Engineering,Guizhou University,Guiyang 550025,China)
出处
《智能计算机与应用》
2021年第6期139-143,共5页
Intelligent Computer and Applications
基金
国家自然科学基金(61663005)。
关键词
人体姿态估计
深度学习
机器视觉
图像处理
人工智能
human pose estimation
deep learning
computer vision
image processing
artificial intelligence