摘要
文章提出基于超像素的三维人体步态运动特征精密提取算法.对三维人体步态运动特征点进行分割,并将分割过程能量函数最小化.利用多元统计模型对分割后的步态运动特征噪音相关信息进行准确估计,设置分形小波的无噪编码,重构三维人体步态运动特征信号,完成三维人体步态运动特征噪声的高效去除.所提算法在三维人体步态运动特征增强和特征提取方面,均优于当前算法,且提取能耗低于当前方法70%左右,可为人体步态运动的进一步研究提供参考依据.
In this artical a new algorithm based on super-pixel is proposed to extract the gait feature of human body. Firstly, the feature points of human gait motion are segmented, and the energy function of the segmentation process is minimized. The multivariate statistical model is used to accurately estimate the noise related information of the segmented gait motion features, and the noiseless coding of fractal wavelet is set to reconstruct the 3D human gait motion feature signals. The feature noise of human gait motion is removed efficiently. The experimental results show that the proposed algorithm is superior to the current algorithm in the enhancement and feature extraction of human gait motion, and the energy consumption of the proposed algorithm is about 70% lower than that of the current method, which can provide a reference for the further study of human gait motion.
作者
李扬中
LI Yang-zhong(South fujian institute of science and technology, Quanzhou 362332 ,Chin)
出处
《微电子学与计算机》
CSCD
北大核心
2018年第7期133-136,共4页
Microelectronics & Computer
关键词
超像素
步态运动特征
精密提取
去噪
super pixel
gait motion characteristics
precision extraction
denoising
