采用自适应变异粒子群优化SVM的行为识别

张国梁; 贾松敏; 张祥银; 徐涛

doi:10.3788/OPE.20172506.1669

您当前的位置：

首页 >

文章列表页 >

采用自适应变异粒子群优化SVM的行为识别

信息科学 | 更新时间：2020-07-07

- 采用自适应变异粒子群优化SVM的行为识别
- Action recognition based on adaptive mutation particle swarm optimization for SVM
- 光学精密工程 2017年25卷第6期页码：1669-1678
- 作者机构：
  
  北京工业大学信息学部, 北京 100124
- 作者简介：
  
  张国梁(1990-), 男, 山西临汾人, 博士研究生, 2015年于太原科技大学获得硕士学位, 主要从事机器人分散控制、人体动作识别及智能设备方面的研究。E-mail:285719262@qq.com ZHANGGuo-liang, E-mail:285719262@qq.com
- 基金信息：
  
  国家自然科学基金资助项目(61175087);北京工业大学智能机器人“大科研”推进计划“助老智能轮椅床自主测控系统的研究与实现”资助项目(040000546317552)
- DOI：10.3788/OPE.20172506.1669
  中图分类号： TP391
- 收稿日期：2016-12-23，
  
  录用日期：2017-1-24，
  
  纸质出版日期：2017-06-25
- 稿件说明：
移动端阅览
张国梁, 贾松敏, 张祥银, 等. 采用自适应变异粒子群优化SVM的行为识别[J]. 光学精密工程, 2017,25(6):1669-1678.

Guo-liang ZHANG, Song-min JIA, Xiang-yin ZHANG, et al. Action recognition based on adaptive mutation particle swarm optimization for SVM[J]. Optics and precision engineering, 2017, 25(6): 1669-1678.
张国梁, 贾松敏, 张祥银, 等. 采用自适应变异粒子群优化SVM的行为识别[J]. 光学精密工程, 2017,25(6):1669-1678. DOI： 10.3788/OPE.20172506.1669.

Guo-liang ZHANG, Song-min JIA, Xiang-yin ZHANG, et al. Action recognition based on adaptive mutation particle swarm optimization for SVM[J]. Optics and precision engineering, 2017, 25(6): 1669-1678. DOI： 10.3788/OPE.20172506.1669.

摘要

为了提高对视频序列中人体行为的识别能力，建立了基于局部特征的动作识别框架。通过时空特征提取及编码和SVM分类器参数优化两部分对该框架所涉及算法进行了研究。首先，采用Harris3D检测器获取时空兴趣点（STIP），以方向梯度直方图（HOG）和光流方向直方图（HOF）对STIP进行描述，并引入Fisher向量实现对特征描述子的编码；由于固定参数下SVM动作分类模型存在泛化能力不足的问题，将粒子群算法应用于各动作分类器参数寻优过程中，针对种群多样性逐代变化的特点，构建粒子聚集度模型，并利用其动态调节各代粒子的变异概率；最后，利用KTH和HMDB51数据集对所提方法进行验证。结果表明，所提自适应变异粒子群算法（AMPSO）能够有效避免种群陷入局部最优，具备较强的全局寻优能力；在KTH和HMDB51数据集上的识别准确率分别为87.50%和26.41%，优于其余2种识别方法。实验证明，AMPSO算法收敛性能良好且整体识别框架具有较高的实用性和准确性。

Abstract

The action recognition framework based on local features was established to improve the recognition ability of human behavior in video sequences. The algorithms related to the framework were researched through spatial temporal features extracting and encoding and parameters optimization of SVM classifier. Firstly

the feature descriptors composed of Histograms of Oriented Gradients (HOG) and Histograms of Optical Flow (HOF) were used to describe Space Time Interest Points (STIP) achieved by the Harris 3D detector and then encoded by Fisher Vector (FV). Due to the generalization ability of Support Vector Machine (SVM) model for action classification under fixed parameters was insufficient

the particle swarm optimization algorithm was applied to the parameter optimization of each action classifier. According to the characteristics of population diversity changed from generation to generation

the constructed particles aggregation degree model was used to adjust mutation probability for each generation of particles dynamically. Finally

the proposed method was verified by KTH and HMDB51 data sets. The results show that the Adaptive Mutation Particle Swarm Optimization (AMPSO) algorithm can avoid the local optimum and has strong global optimization capability. The recognition accuracies on KTH and HMDB51 data sets are 87.50% and 26.41%

respectively

which are better than two recognition methods. Experimental results indicate that the AMPSO algorithm has good convergence performance and the overall recognition framework has high practicability and accuracy.

关键词

Keywords

references

王世刚, 鲁奉军, 赵文婷, 等.应用在线随机森林投票的动作识别[J].光学精密工程, 2016, 24(8): 2010-2017.

WANG S G, LU F J, ZHAO W T, et al.. Action recognition based on-line random forest voting [J]. Opt. Precision Eng., 2016, 24(08): 2010-2017. (in Chinese)

PENG X J, WANG L M, WANG X X, et al.. Bag of visual words and fusion methods for action recognition: Comprehensive study and good practice [J]. Computer Vision and Image Understanding, 2016, 150: 109-125.

SCHULDT C, LAPTEV I, CAPUTO B. Recognizing human actions: a local SVM approach [C]. 2004 IEEE Proceedings of the 17th International Conference on Pattern Recognition, Cambridge, UK, ICPR, 2004: 32-36.

LAPTEV I. On space-time interest points [J]. International Journal of Computer Vision, 2005, 64(2-3): 107-123.

WANG H, SCHMID C. Action recognition with improved trajectories [C]. 2013 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, Portland, USA, CVPR, 2013: 3551-3558.

LIU L, WANG L, LIU X. In defense of soft-assignment coding [C]. 2011 International Conference on Computer Vision, Barcelona, Spain, ICCV, 2011: 2486-2493.

PERRONNIN F, SANCHEZ J, MENSINK T. Improving the fisher kernel for large-scale image classification [C]. 2010 European conference on computer vision, Crete, Greece, ECCV, 2010: 143-156.

宋悦, 林志贤, 姚剑敏.基于ε-支持向量回归机的背光源亮度均匀性评价方法[J].液晶与显示, 2015, 30(05): 857-863.

SONG Y, LIN Z X, YAO J M. Luminance uniformity evaluation for backlight based on ε-support vector regression [J]. Chinese Journal of Liquid Crystals and Displays, 2015, 30(05): 857-863. (in Chinese)

SHEN L, CHEN H, YU Z, et al.. Evolving support vector machines using fruit fly optimization for medical data classification [J]. Knowledge-Based Systems, 2016, 96: 61-75.

HUANG C L, WANG C J. A GA-based feature selection and parameters optimizationfor support vector machines [J]. Expert Systems with Applications, 2006, 31(2): 231-240.

WANG G P, CHEN S Y, LIU J.. A Short-Term prediction model based on support vector regression optimized by artificial Fish-Swarm algorithm [J]. International Journal of Control and Automation, 2015, 8(7): 237-250.

段海滨, 张祥银, 徐春芳.仿生智能计算[M].北京:科学出版社, 2011.

DUAN H B, ZHANG X Y, XU C F. Bio-inspired Computing [M]. Beijing: Science Press, 2011. (in Chinese)

刘志强, 尹建芹, 张玲, 等.基于Kinect数据主成分分析的人体动作识别[J].光学精密工程, 2015, 23(10): 702-711.

LIU Z Q, YIN J Q, ZHANG L, et al.. Human action recognition based on Kinect data principal component analysis [J]. Opt. Precision Eng., 2015, 23(10): 702-711. (in Chinese)

于海琦, 刘真.一种基于GA-BP神经网络结合PCA的LCD显示器光谱特征化模型[J].液晶与显示, 2016, (02): 201-207.

YU H Q, LIU Z. Spectral characterization model of liquid crystal display based on GA-BP network and PCA [J]. Chinese Journal of Liquid Crystals and Displays, 2016, 31(02): 201-207. (in Chinese)

SANCHEZ J, PERRONNIN F, MENSINK T, et al.. Image classification with the fisher vector: Theory and practice [J]. International Journal of Computer Vision, 2013, 105(3): 222-245.

ABOUHAWWASH M, SEADA H, DEB K. Towards faster convergence of evolutionary multi-criterion optimization algorithms using karush kuhn tucker optimality based local search [J]. Computers & Operations Research, 2017, 79(1):331-346.

WANG L, QIAO Y, TANG X. Action recognition with trajectory-pooled deep-convolutional descriptors [C]. 2015 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, Boston, USA, CVPR, 2015: 4305-4314.

LAN Z, LIN M, LI X, et al.. Beyond gaussian pyramid: Multi-skip feature stacking for action recognition [C]. 2015 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, Boston, USA, CVPR, 2015: 204-212.

KUEHNE H, JHUANG H, STIEFELHAGEN R. HMDB51: A large video database for human motion recognition [J]. High Performance Computing in Science and Engineering, 2013, 12(1): 571-582.

FERNANDO B, GAVVES S, MOGROVEJO O, et al.. Rank pooling for action recognition [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(4):773-787.

王世刚, 孙爱朦.基于时空兴趣点的单人行为及交互行为识别[J].吉林大学学报(工学版), 2015, 45(01): 304-308.

WANG S G, SUN A M. Single and interactive human behavior recognition algorithm based on spatio-temporal interest point [J]. Journal of Jilin University (Engineering and Technology Edition), 2015, 45(01): 304-308. (in Chinese)

浏览量

183

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于改进SIFT特征提取的车标识别

结合形态学属性断面与支持向量机的合成孔径雷达图像变化检测

基于改进支持向量机的目标威胁估计

采用降维技术的红外目标检测与识别

联合梯度直方图和局部二值模式特征的人体检测