基于深度图像的人体行为识别

doi:10.16182/j.issn1004731x.joss.201805003

系统仿真学报 ›› 2018, Vol. 30 ›› Issue (5): 1641-1649.doi: 10.16182/j.issn1004731x.joss.201805003

基于深度图像的人体行为识别

唐超¹, 张苗辉^2,*, 李伟³, 曹峰⁴, 王晓峰¹, 童晓红⁵

1. 合肥学院计算机科学与技术系,合肥 230601;
2. 江西省科学院能源研究所,南昌 330096;
3. 厦门理工学院计算机与信息工程学院,厦门 361024;
4. 山西大学计算机与信息技术学院,太原 030006;
5. 合肥职业技术学院信息中心,合肥 238000

收稿日期:2017-08-19 修回日期:2017-09-18 出版日期:2018-05-08 发布日期:2019-01-03
作者简介:唐超(1977-),男,安徽合肥,博士,讲师,研究方向为机器学习和计算机视觉。
基金资助:
国家自然科学基金(61672204, 41401521, 61602220),合肥学院人才科研基金(15RC07),安徽高校优秀拔尖人才培育资金(gxfx2017099),山西省青年科技研究基金(2015021101),福建省自然科学基金(2016J01325, 2015J05015),江西省自然科学基金(20161BAB21057)

Human Action Recognition Based on Depth Image

Tang Chao¹, Zhang Miaohui^2,*, Li Wei³, Cao Feng⁴, Wang Xiaofeng¹, Tong Xiaohong⁵

1. Department of Computer Scince and Technology, Hefei University, Hefei 230601, China;
2. Energy Research Institute, Jiangxi Academy of Sciences, Nanchang 330096, China;
3. School of Computer and Information Engineering, Xiamen University of Technology, Xiamen 361024, China;
4. School of Computer and Information Science, Shanxi University, Taiyuan 030006, China;
5. Information center, Hefei Technology College, Hefei 238000, China

Received:2017-08-19 Revised:2017-09-18 Online:2018-05-08 Published:2019-01-03

摘要/Abstract

摘要： 由于人体动作的复杂性和非刚性特点,传统的基于RGB视频数据的人体行为识别是一个非常具有挑战性的研究课题。针对现有基于RGB视频数据识别方法的不足,提出了一种基于深度图像数据的人体行为识别方法,该方法将深度差值运动历史图像中分块均值特征与Gabor特征进行融合,采用泛化能力较好的旋转森林算法进行建模。在DHA深度数据集上实验结果表明,相比其它监督行为识别分类算法,基于深度图像的方法具有简单、快速,高效的特点。

关键词: 人体行为识别, 深度图像, 深度差值运动历史图像, Gabor特征, 旋转森林

Abstract: Because of the complexity and non-rigidity of human actions, traditional human action recognition based on RGB video data is a very challenging research topic. According to some deficiencies of existing recognition method based on RGB video data, a novel human action recognition method is proposed based on depth image data. In this new method, the block mean feature in the depth difference motion historical image is fused with the Gabor feature as mixed features and then a rotation forest algorithm is used to model. The experimental results show that the proposed method is simple, fast and efficient compared with other supervised action recognition algorithms on DHA depth datasets.

Key words: human action recognition, depth image, depth difference motion historical image, Gabor feature, rotation forest

中图分类号:

TP391

唐超, 张苗辉, 李伟, 曹峰, 王晓峰, 童晓红. 基于深度图像的人体行为识别[J]. 系统仿真学报, 2018, 30(5): 1641-1649.

Tang Chao, Zhang Miaohui, Li Wei, Cao Feng, Wang Xiaofeng, Tong Xiaohong. Human Action Recognition Based on Depth Image[J]. Journal of System Simulation, 2018, 30(5): 1641-1649.

参考文献

[1] LO PRESTI L, LA CASCIA M.3D skeleton-based human action classification: A survey[J]. Pattern Recognition (S0031-3203), 2016, 53: 130-147.
[2] DAS DAWN D, SHAIKH S H.A comprehensive survey of human action recognition with spatio-temporal interest point (STIP) detector[J]. Visual Computer (S0178-2789), 2016, 32(3): 289-306.
[3] ZIAEEFARD M, BERGEVIN R.Semantic human activity recognition: A literature review[J]. Pattern Recognition (S0031-3203), 2015, 48(8): 2329-2345.
[4] LIM C H, VATS E, CHAN C S.Fuzzy human motion analysis: A review[J]. Pattern Recognition (S0031-3203), 2015, 48(5): 1773-1796.
[5] CHEN C, JAFARI R, KEHTARNAVAZ N.A survey of depth and inertial sensor fusion for human action recognition[J]. Multimedia Tools and Applications (S1380-7501), 2017, 76(3): 4405-4425.
[6] CAI Z, HAN J, LIU L, et al.RGB-D datasets using microsoft kinect or similar sensors: a survey[J]. Multimedia Tools and Applications (S1380-7501), 2017, 76(3): 4313-4355.
[7] ZHANG J, LI W, OGUNBONA P O, et al.RGB-D-based action recognition datasets: A survey[J].Pattern Recognition (S0031-3203), 2016, 60: 86-105.
[8] ZHANG B, YANG Y, CHEN C, et al.Action Recognition Using 3D Histograms of Texture and A Multi-Class Boosting Classifier[J]. IEEE Transactions on Image Processing (S1057-7149), 2017, 26(10): 4648-4660.
[9] ZHANG H, ZHONG P, HE J, et al.Combining depth-skeleton feature with sparse coding for action recognition[J]. Neurocomputing (S0925-2312), 2017, 230: 417-426.
[10] YANG Y, DENG C, GAO S, et al.Discriminative Multi-instance Multitask Learning for 3D Action Recognition[J]. IEEE Transactions on Multimedia (S1520-9210), 2017, 19(3): 519-529.
[11] UDDIN M Z.Human activity recognition using segmented body part and body joint features with hidden Markov models[J]. Multimedia Tools and Applications (S1380-7501), 2017, 76(11): 13585-13614.
[12] RAHMANI H, MAHMOOD A, DU H, et al.Histogram of Oriented Principal Components for Cross-View Action Recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2016, 38(12): 2430-2443.
[13] AKAGUNDUZ E, ASLAN M, SENGUR A, et al.Silhouette Orientation Volumes for Efficient Fall Detection in Depth Videos[J]. IEEE Journal of Biomedical and Health Informatics (S2168-2194), 2017, 21(3): 756-763.
[14] YANG J, YANG J-Y, ZHANG D, et al.Feature fusion: Parallel strategy vs. serial strategy[J]. Pattern Recognition (S0031-3203), 2003, 36(6): 1369-1381.
[15] LLINAS J, HALL D L.Introduction to multi-sensor data fusion[C]// Proc of the 1998 IEEE International Symposium on Circuits and Systems. NEW YORK, NY: IEEE, 2002: 537-540.
[16] RODRIGUEZ J J, KUNCHEVA L I.Rotation forest: A new classifier ensemble method[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2006, 28(10): 1619-1630.
[17] LIN Y C, HU M C, CHENG W H, et al.Human action recognition and retrieval using sole depth information[C]// Proc of the 20th ACM International Conference on Multimedia. Nara, Japan, 2012: 1053-1056.

基于深度图像的人体行为识别

Human Action Recognition Based on Depth Image

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 7

编辑推荐

Metrics

本文评价

[1]	阴敬方, 朱登明, 石敏, 王兆其. 基于引导对抗网络的人体深度图像修补方法[J]. 系统仿真学报, 2020, 32(7): 1312-1321.
[2]	陈莹, 沈栎. 基于特征回归的单目深度图无标记人体姿态估计[J]. 系统仿真学报, 2020, 32(2): 269-277.
[3]	唐超, 张苗辉, 李伟, 曹峰, 王晓峰, 童晓红. 融合局部与全局特征的人体动作识别[J]. 系统仿真学报, 2018, 30(7): 2497-2506.
[4]	王松, 刘复昌, 黄骥, 许威威, 董洪伟. 基于卷积神经网络的深度图姿态估计算法研究[J]. 系统仿真学报, 2017, 29(11): 2618-2623.
[5]	杨家博, 杨刚, 杨猛. 基于深度图像分割与物体跟踪的增强现实系统[J]. 系统仿真学报, 2017, 29(11): 2788-2795.
[6]	叶继华, 兰清平, 刘长红, 王仕民. 结合互信息量和Log-Gabor特征的嵌入式人脸识别[J]. 系统仿真学报, 2016, 28(9): 2214-2219.
[7]	王满一, 宋亚玲, 李玉, 张良. 结合区域光流特征的时序模板行为识别[J]. 系统仿真学报, 2015, 27(5): 1146-1151.