摘要
3D手势姿态估计是计算机视觉领域一个重要的研究方向,在虚拟现实、增强现实、人机交互、手语理解等领域中具有重要的研究意义和广泛的应用前景.深度学习技术已经广泛应用于3D手势姿态估计任务并取得了重要研究成果,其中深度图像具有的深度信息可以很好地表示手势纹理特征,深度图像已成为手势姿态估计任务重要数据源.本文首先全面阐述了手势姿态估计发展历程、常用数据集、数据集标记方式和评价指标;接着根据深度图像的不同展现形式,将基于深度图像的数据驱动手势姿态估计方法分为基于简单2D深度图像、基于3D体素数据和基于3D点云数据,并对每类方法的代表性算法进行了概括与总结;最后对手势姿态估计未来发展进行了展望.
3D hand pose estimation is an important research direction in the field of computer vision,which has essencial research significance and wide application prospects in the fields of virtual reality,augmented reality,human-computer interaction and sign language understanding.Deep learning has been widely used in 3D hand pose estimation tasks and has achieved considerable results.Among them,the depth information contained in the depth image can well represent the texture characteristics of the hand poses,and the depth image has become an important data source for hand pose estimation tasks.Firstly,development history benchmark data sets,marking methods and evaluation metrics of hand pose estimation were introduced.After that,according to the different presentation forms of depth maps,the data-driven hand pose estimation methods based on depth images are divided into simple 2D depth map based methods,3D voxel data based methods and 3D point cloud data based method,and we further analyzed and summarized the representative algorithms of them.At the end of this paper,we discussed the development trend of hand pose estimation in the future.
作者
王丽萍
汪成
邱飞岳
章国道
WANG Li-ping;WANG Cheng;QIU Fei-yue;ZHANG Guo-dao(College of Computer Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China;College of Education Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2021年第6期1227-1235,共9页
Journal of Chinese Computer Systems
基金
浙江省重点研发计划基金项目(2018C01080)资助.
关键词
3D手势姿态估计
深度学习
深度图像
虚拟现实
人机交互
3D hand pose estimation
deep learning
depth map
virtual reality
human-computer interaction
