| 摘 要: 随着二维人体关键点检测应用领域的扩展,对二维人体关键点检测的检测速度和精度提出了更高的要求。为了进一步分析二维人体关键点检测,首先介绍了基于传统方法和基于深度方法的二维人体关键点检测的模型,其次介绍了二维人体关键点检测的常用数据集和研究方向,最后得出了基于深度学习的二维人体关键点检测可以高精度地定位人体关键点的结论。未来,可以将二维人体关键点检测应用到更多新的领域进行更深入的研究。同时,提高模型的鲁棒性和泛化能力,探索更加高效的模型架构,以及减少模型复杂度等将成为研究热点。 |
| 关键词: 二维图像;关键点检测;神经网络;深度学习 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 国家自然科学基金(82127807);上海市分子影像学重点实验室建设项目(18DZ2260400). |
|
| Overview of Key Point Detection Algorithms for 2D Human Body |
|
SHI Xiaoqiang1, HUANG Gang1,2, SU Keyi3
|
(1. School of Health Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China;
2. Shanghai Key Laboratory of Molecular Imaging, Jiading District Central Hospital, Shanghai University of Medicineand Health Sciences, Shanghai 201318, China;
3. China State Institute of Pharmaceutical Industry, Shanghai 201203, China)
858961788@qq.com; huanggang@sumhs.edu.cn; 310460791@qq.com
|
| Abstract: As the application field of 2D human body key point detection expands, higher requirements are put forward for the detection speed and accuracy of 2D human body key point detection. In order to further analyze 2D human body key point detection, this paper first introduces 2D human body key point detection models based on traditional methods and deep learning methods. Then, commonly used datasets and research directions for 2D human body key point detection are introduced. Finally, it is concluded that 2D human body key point detection based on deep learning can accurately locate human body key points. In future, 2D human body key point detection can be applied to more new fields for further in-depth research. At the same time, improving the robustness and generalization ability of the model, exploring more efficient model architectures, and reducing model complexity will become research focuses. |
| Keywords: 2D image; key point detection; neural network; deep learning |
