摘要
由于人体运动过程中行为的多样性以及复杂性,导致人体行为特征方向估计模型的效果不理想、准确率偏低等问题。为此,构建基于大数据的人体行为特征方向估计模型。通过人体运动图像中的深度信息分别计算出不同像素点在水平方向和垂直方向的梯度值,再计算不同像素点与邻域像素点之间的差值,获取人体行为特征。对图像中的关键参数进行自适应处理,利用遗传算法对关键参数进行寻优,并构建基于大数据的人体行为特征方向估计模型。实验结果表明,与传统的人体行为特征方向估计模型相比,所提估计模型在人体行为特征方向估计效果以及准确率方面都有较大幅度的提升。
Due to the diversity and complexity of behavior during human motion,the estimation model of human behavior features is not ideal and the accuracy is low.Therefore,a model to estimate human behavior feature direction based on big data was constructed.The gradient values of different pixel points on the horizontal direction and the vertical direction are respectively computed by the depth information in human motion image,and the difference between different pixel points and adjacent pixel points was calculated to obtain the human behavior feature.The key parameters in image were adaptively processed.Finally,the genetic algorithm was used to optimize key parameters,and the estimation model of human behavior feature direction based on big data was constructed.Following conclusion can be drawn from Simulation results show that,compared with the traditional estimation model of human behavior feature direction,the proposed estimation model has a great improvement in the estimation effect and accuracy of human behavior feature.
作者
刘静
LIU Jing(Youth College of Politics Science of Inner Mongolia Normal University,Hohhot Inner Mongolia 010051,China)
出处
《计算机仿真》
北大核心
2019年第9期422-425,451,共5页
Computer Simulation
关键词
大数据
人体行为
特征方向
估计模型
Big data
Human behavior
Feature direction
Estimation model
作者简介
刘静(1981-),女(满族),内蒙古乌兰察布人,硕士,讲师,研究方向:计算机应用。
