基于卷积神经网络的时空权重姿态运动特征提取算法

doi:10.3969/j.issn.0255-8297.2021.04.007

摘要/Abstract

摘要： 在传统姿态运动特征提取过程中存在有效提取效率低的问题，于是提出了基于卷积神经网络（convolutional neural network，CNN）算法的时空权重姿态运动特征提取算法。针对所选择的运动时空样本，提取相应的时空运动关键帧并以静态图像的形式输出；采取运动目标检测、图像增强等多项措施完成初始运动图像的预处理工作；借助CNN将运动特征矢量化；采用时空权重自适应插值方法减少运动边缘检测误差，从姿态边缘特征和姿态运动时空特征两方面实现姿态运动特征提取，并输出提取结果。与传统算法进行对比实验的结果表明，所提出的算法在有效特征数量方面得到了提升。

关键词: 卷积神经网络, 时空权重, 运动姿态, 运动特征, 特征提取算法

Abstract: In traditional attitude motion feature extraction process, there is the problem of low efficiency. As to this, a temporal and spatial weight attitude motion feature extraction algorithm based on convolutional neural network (CNN) algorithm is proposed in this paper. First, from selected motion spatio-temporal samples, corresponding spatio-temporal motion keyframes are extracted and output in the form of static images. Second, initial moving images are preprocessed by using moving object detection, image enhancement and other measures. Then the motion feature is vectorized by CNN, and the adaptive interpolation method of space-time weight is used to reduce the error of motion edge detection. Finally, feature extraction of attitude motion is realized from two aspects of attitude edge feature and space-time feature of attitude motion, and extraction results are output. Compared with the traditional algorithm, experimental results show that the proposed algorithm improves the number of effective features.

Key words: convolutional neural network (CNN), weight of time and space, movement posture, movement characteristic, feature extraction algorithm

中图分类号:

TP391

郑长亮, 庞明. 基于卷积神经网络的时空权重姿态运动特征提取算法[J]. 应用科学学报, 2021, 39(4): 594-604.

ZHENG Changliang, PANG Ming. Spatial-Temporal Weight Attitude Motion Feature Extraction Algorithm Using Convolutional Neural Network[J]. Journal of Applied Sciences, 2021, 39(4): 594-604.

参考文献

[1] 肖铭杰, 路强, 陆亦扬, 等. 基于地形特征重要度的地形网格简化与调整[J]. 合肥工业大学学报(自然科学版), 2019, 2(5):651-654. Xiao M J, Lu Q, Lu Y Y, et al. Simplification and adjustment of terrain grid based on terrain feature importance[J]. Journal of Hefei University of Technology (Natural Science Edition), 2019, 2(5):651-654. (in Chinese)
[2] 陈华伟, 袁小翠, 吴禄慎, 等. 基于曲率突变分析的点云特征线自动提取算法[J]. 光学精密工程, 2019, 41(5):1218-1228. (in Chinese) Chen H W, Yuan X C, Wu L S, et al. Automatic feature extraction algorithm of point cloud based on curvature mutation analysis[J]. Optical Precision Engineering, 2019, 41(5):1218-1228. (in Chinese)
[3] 许春冬, 周静, 应冬文, 等. 心音信号MFCC特征向量提取方法的优化[J]. 信号处理, 2019, 9(3):410-418. Xu C D, Zhou J, Ying D W, et al. Optimization of MFCC feature vector extraction method for heart sound signal[J]. Signal Processing, 2019, 9(3):410-418. (in Chinese)
[4] 梁令羽, 张天天, 何为. 多尺度卷积神经网络的头部姿态估计[J]. 激光与光电子学进展, 2019, 56(13):71-78. Liang L Y, Zhang T T, He W. Head pose estimation based on multiscale convolutional neural network[J]. Progress in Laser and Optoelectronics, 2019, 56(13):71-78. (in Chinese)
[5] 范书瑞, 贾雅亭, 刘晶花. 基于三轴加速度传感器人体姿态识别的特征选择[J]. 应用科学学报, 2019, 37(3):427-436. Fan S R, Jia Y T, Liu J H. Feature selection of human activity recognition based on tri-axial accelerometer[J]. Journal of Applied Sciences, 2019, 37(3):427-436. (in Chinese)
[6] Qiu L, Sang D W, Fengy H, et al. Analysis of particle motion characteristics in a funnel-shape moving bed[J]. Journal of Engineering Thermophysics, 2018, 39(12):2708-2713.
[7] Xin Z, Xi K, Dai J S. Gait planning and motion characteristic analysis of a metamorphic quadruped walking robot[J]. Journal of Central South University, 2018, 3(2):265-270.
[8] 陈国军, 杨静, 程琰, 等. 基于RGBD的实时头部姿态估计[J]. 图学学报, 2019, 40(4):681-688. Chen G J, Yang J, Cheng Y, et al. Real time head pose estimation based on RGBD[J]. Journal of Graphics, 2019, 40(4):681-688. (in Chinese)
[9] Lin Y N, Yeh T, Ma K, et al. Source characteristics of the 2016 Meinong (ML 6.6), Taiwan, earthquake, revealed from dense seismic arrays:double sources and pulse velocity ground motion[J]. Bulletin of the Seismological Society of America, 2018, 108(1):514-520.
[10] Wang G X, Li A J, Yang G Y, et al. Fractional order attitude stability control for subsatellite of tethered satellite system during deployment[J]. Applied Mathematical Modelling, 2018, 62(6):785-787.
[11] 王容川, 庄志洪, 王宏波, 等. 基于卷积神经网络的雷达目标HRRP分类识别方法[J]. 现代雷达, 2019, 23(5):33-38. Wang R C, Zhuang Z H, Wang H B, et al. HRRP classification and recognition method of radar target based on convolutional neural network[J]. Modern Radar, 2019, 23(5):33-38. (in Chinese)
[12] 成翔昊, 达飞鹏, 邓星. 基于关键点的由粗到精三维人脸特征点定位[J]. 仪器仪表学报, 2018, 5(10):256-264. Cheng X H, Da F P, Deng X. Coarse to fine 3D facial feature point localization based on key points[J]. Journal of Instrumentation, 2018, 5(10):256-264. (in Chinese)
[13] Huang Y, Zhang Y G, Chang L B. A new fast in-motion coarse alignment method for GPS-aided low-cost SINS[J]. IEEE/ASME Transactions on Mechatronics, 2018, 23(3):1303-1313.
[14] 蔡冠蓝. 柔性姿态估计和时空特征结合的乒乓球动作视频片段关键帧提取[J]. 科学技术与工程, 2019, 19(25):25-29. Cai G L. Key frame extraction of table tennis action video clip based on flexible pose estimation and spatiotemporal characteristics[J]. Science, Technology and Engineering, 2019, 19(25):25-29. (in Chinese)
[15] Sun Z P, Lu Y J, Xu L X, et al. Bionic attitude transformation combined with closed motion for a free floating space robot[J]. Journal of Beijing Institute of Technology, 2018, 27(1):118-126.
[16] Sun Q Q, Liang F W, Feng W, et al. Investigation on the geometrical characteristics of secondary arc by image edge detection[J]. IEEE Transactions on Plasma Science, 2018, 46(99):2016-2025
[17] 贾迪, 李玉秀, 赵明远, 等. 面向像对直线特征匹配的线特征矫正与提纯方法[J]. 中国图象图形学报, 2019, 3(7):1176-1187. Jia D, Li Y X, Zhao M Y, et al. Line feature correction and purification method for image to line feature matching[J]. Chinese Journal of Image Graphics, 2019, 3(7):1176-1187. (in Chinese)
[18] Yan L, Zheng Y. Evolutionary spatiotemporal community discovery in dynamic weighted networks[J]. Journal of Internet Technology, 2018, 19(2):499-506.
[19] 易中贵, 戈新生. 间接Legendre伪谱法的欠驱动航天器姿态运动轨迹跟踪[J]. 宇航学报, 2018, 39(6):648-655. Yi Z G, Ge X S. Attitude tracking of underactuated spacecraft using indirect Legendre pseudospectral method[J]. Acta Astronautica Sinica, 2018, 39(6):648-655. (in Chinese)
[20] 崔景春, 王静. 基于增强头部姿态估计的人脸表情识别模型[J]. 计算机科学, 2019, 12(6):322-327. Cui J C, Wang J. Facial expression recognition model based on enhanced head pose estimation[J]. Computer Science, 2019, 12(6):322-327. (in Chinese)