计算机应用与软件2017，Vol.34Issue(6)：176-179,200,5.DOI:10.3969/j.issn.1000-386x.2017.06.032
基于关节间夹角变化序列的动作识别算法研究
RESEARCH ON ACTION RECOGNITION ALGORITHM BASED ON THE JOINT ANGLE CHANGE SEQUENCE
摘要
Abstract
Aiming at the problem of complex background and viewpoint change in human action recognition, an action recognition algorithm based on the joint angle change sequence is proposed.Inspired by the mechanism and robotics, the motion of the human body is represented by the sequence of the angle between the adjacent joints and the angle between the non-adjacent joints.Then the KNN classifier is used to classify the action.Because the duration of action of different individuals is not the same, the DTW algorithm is used to calculate the distance between the action samples.Finally, the proposed method is validated on open datasets UTD-MHAD and KARD, and the experimental results show the effectiveness of the method.关键词
K近邻/动态时间规整/动作识别Key words
KNN/DTW/Action recognition分类
信息技术与安全科学引用本文复制引用
王松,党建武,王阳萍,杜晓刚..基于关节间夹角变化序列的动作识别算法研究[J].计算机应用与软件,2017,34(6):176-179,200,5.基金项目
国家自然科学基金项目(61562057) （61562057）
甘肃自然科学基金项目(145RJZA080) （145RJZA080）
兰州交通大学校青年基金项目(2013009,2013005). （2013009,2013005）
