[1]梅雪,胡石,许松松,等.基于多尺度特征的双层隐马尔可夫模型及其在行为识别中的应用[J].智能系统学报,2012,7(06):512-517.
　MEI Xue,HU Shi,XU Songsong,et al.Multi scale feature based double layer HMM and its application in behavior recognition[J].CAAI Transactions on Intelligent Systems,2012,7(06):512-517.
点击复制

基于多尺度特征的双层隐马尔可夫模型及其在行为识别中的应用(/HTML)
分享到：

《智能系统学报》[ISSN:1673-4785/CN:23-1538/TP]

卷:
第7卷
期数:
2012年06期
页码:
512-517
栏目:
出版日期:
2012-12-25

文章信息/Info

Title:
Multi scale feature based double layer HMM and its application in behavior recognition
文章编号:
1673-4785(2012)06-0512-06
作者:
梅雪; 胡石; 许松松; 张继法
南京工业大学 自动化与电气工程学院，江苏 南京 211816
Author(s):
MEI Xue;  HU Shi;  XU Songsong;  ZHANG Jifa
College of Automation and Electrical Engineering, Nanjing University of Technology, Nanjing 211816, China
关键词:
双层隐马尔可夫模型; 行为识别; 多尺度特征; 智能视频监控
Keywords:
doublelayer HMM (DLHMM);  behavior recognition;  multiscale feature;  intelligent video surveillance
分类号:
TP391.4
文献标志码:
A
摘要:
借鉴人类视觉感知所具有的多尺度、多分辨性的特性，针对智能视频监控系统的人体运动行为识别，提出了一种基于多尺度特征的双层隐马尔可夫模型.根据人体行为关键姿态数确定HMM的状态数目，发掘人体运动行为隐藏的多尺度结构间的关系，将运动轨迹和人体姿态边缘小波矩2个不同尺度特征应用于2层HMM，提供更为丰富的行为尺度间的相关信息.分别用Weizmann人体行为数据库和自行拍摄的室内视频，对人体运动行为识别进行仿真实验，结果表明，五状态HMM模型更符合人体运动行为特点，基于多尺度特征的五状态双层隐马尔可夫模型具有较高的识别率.
Abstract:
Learning from multiscale and multidistinguish attributes of human beings’ visual perception and aiming at human movement behavior recognition in intelligent video surveillance system, a doublelayer hidden markov model (DLHMM) is developed based on multiscale behavior features. Considering the human behavior characteristics, the number of HMM states is according to the number of key gestures selected. Discovering the relationship between the multiscale structures hidden in the human movement behavior, two different scale featureshuman motion trajectory and wavelet moment of human gesture’s edge, are applied respectively in two layers of DLHMM, so as to provide more scale information about behavior. Experiments, using Israel Weizmann human behavior database and human actions indoor recorded by ourselves, show the fivestate HMM more accords with the human motion behavior characteristics, and the fivestate DLHMM based on multiscale feature has a higher recognition rate compared with traditional methods using one layer HMM.

参考文献/References:

［1］黎洪松，李达.人体运动分析研究的若干新进展［J］. 模式识别与人工智能, 2009, 22(1): 7078. 
LI Hongsong， LI Da.Some advances in human motion analysis［J］.Pattern Recognition and Aritificial Intelligence, 2009, 22(1): 7078.
［2］ WEINLAND D, RONFARD R. A survey of vision based methods for action representation, segmentation and recognition［J］. Computer Vision and Image Understanding, 2011, 115(2): 224241.
［3］ PAUCHOO C, De CHIN L. A daily behavior enabled hidden Markov model for human behavior understanding［J］. Pattern Recognition, 2008, 41(2): 15721580.
［4］ NASCIMENTO C J, FIGUEIREDO A M T, MARQUES S J. Trajectory classification using switched dynamical hidded Markov models［J］. IEEE Trans on Image Processing, 2010, 19 (5): 13381348.
［5］ HERVIEU A, BOUTHEMY P, CADRE J P L. A HMM based method for recognizing dynamic video contents from trajectories［C］//Proceedings of International Conference on Image Processing. San Antonio, USA, 2007: 533536.
［6］ UDDIN Z, NGUYEN T, JEONG K, et al. Human activity recognition using body jointangle features and hidden Markov model［J］. Electronics and Telecommunications Research Institute, 2011, 33(4): 569579.
［7］ YOUNG L, SUNG C. Activity recognition using hierachical hidden Markov models on a smartphone with 3D accelerometer［C］//6th International Conference on HAIS.Berlin:SpringerVerlag， 2011: 460467.
［8］钱堃，马旭东，戴先中. 基于抽象隐马尔可夫模型的运动行为识别方法［J］. 模式识别与人工智能, 2009, 22(3): 433439.
QIAN Kun, MA Xudong,DAI Xianzhong. Motion activity recognition based on abstract hidden Markov model［J］.Pattern Recognition and Aritificial Intelligence, 2009, 22(3): 433439.
［9］CHEN Changhong, LIANG Jimin, ZHU Xiuchang. Gait recognition based on improved dynamic Bayesian networks［J］. Pattern Recognition, 2011, 44(4): 988995. 
［10］胡石，梅雪. 人体行为动作的形状轮廓特征提取及识别［J］. 计算机工程, 2012, 38(4)： 266268.
HU Shi,MEI Xue.Shape contour feature extraction and recognition of human behavior motion［J］.Computer Engineering, 2012, 38(4)： 266268.
［11］ NEGIN N, FARBOD R, HOSSEIN S. A novel approach to HMMbased speech recognition systems using particle swarm optimization［J］. Mathematical and Computer Modelling, 2010, 52 (12): 19101920.

相似文献/References:

[1]姬晓飞,谢旋,任艳.深度学习的双人交互行为识别与预测算法研究[J].智能系统学报,2020,15(3):484.[doi:10.11992/tis.201812029]
　JI Xiaofei,XIE Xuan,REN Yan.Human interaction recognition and prediction algorithm based on deep learning[J].CAAI Transactions on Intelligent Systems,2020,15(06):484.[doi:10.11992/tis.201812029]
[2]刘董经典,孟雪纯,张紫欣,等.一种基于2D时空信息提取的行为识别算法[J].智能系统学报,2020,15(5):900.[doi:10.11992/tis.201906054]
　LIU Dongjingdian,MENG Xuechun,ZHANG Zixin,et al.A behavioral recognition algorithm based on 2D spatiotemporal information extraction[J].CAAI Transactions on Intelligent Systems,2020,15(06):900.[doi:10.11992/tis.201906054]

备注/Memo

备注/Memo:
收稿日期： 2012-03-15.
网络出版日期：2012-11-16.
基金项目：江苏省高校自然科学基金资助项目（09KJB510002）; 江苏省博士后科研资助计划资助项目（1001027B）;南京工业大学青年学科基金资助项目（39710006）.
通信作者：梅雪.
E-mail: mx@njut.edu.cn.
作者简介：
梅雪,女，1975年生，副教授，硕士生导师，主要研究方向为图像处理、模式识别及计算机视觉.
胡石，男，1988年生，硕士研究生，主要研究领域为模式识别、图像处理.
张继法，男，1987年生，硕士研究生，主要研究领域为模式识别、图像处理.
更新日期/Last Update: 2013-03-19