ZHAN Yin-wei, YU Zhi-zhi, CAI Jun. Posture recognition algorithm based on angle measurement with Kinect[J]. 2014, 33(7): 129-132. DOI: 10.13873/j.1000-97872014.07.025.
基于Kinect角度测量的姿势识别算法
摘要
人体姿势识别在医疗康复等领域中有着重要的作用。为了提高姿势识别的准确率
提出了一种利用Kinect测量角度的人体姿势识别方法。首先
利用Kinect获取人体关节点的空间坐标。然后
通过两点法计算角度的大小并定义人体姿势库。最后
与姿势库进行角度匹配实现姿势识别。实验结果表明:此方法能够实时准确地识别人体姿势
并且可以通过组建不同的姿势库来识别更多的姿势。
Abstract
Posture recognition plays an important role in many applications
such as rehabilitation.To improve the accuracy of posture recognition
propose a new method for human posture recognition based on Kinect.Firstly
Kinect is used to obtain the world coordinates of joints.Then through two-point method to calculate angle and define human posture library.Finally
angle matching is carried out with posture library to realize gesture recognition.Experimental results show that this method can real-time accurately identify postures
and build different posture library to recognize more postures.
关键词
Keywords
references
Human skeleton tracking from depth data using geodesic distances and optical flow [J] . Loren Arthur Schwarz,Artashes Mkhitaryan,Diana Mateus,Nassir Navab. Image and Vision Computing . 2011 (3)
A study to evaluate a low cost virtual reality system for home based rehabilitation of the upper limb following stroke [J] . Penny J. Standen,David J. Brown,Steven Battersby,Marion Walker,Louise Connell,Andy Richardson,Fran Platts,Kate Threapleton,Andy Burton. International Journal on Disability and Human Development . 2011 (4)
Inside the race to hack the Kinect [J] . Jim Giles. New Scientist . 2010 (2789)