计算机应用 ›› 2013, Vol. 33 ›› Issue (04): 957-959.DOI: 10.3724/SP.J.1087.2013.00957

• 人工智能 • 上一篇    下一篇

基于分层分数条件随机场的行为识别

王科俊,吕卓纹,孙国振,阎涛   

  1. 哈尔滨工程大学 自动化学院,哈尔滨 150001
  • 收稿日期:2012-10-16 修回日期:2012-12-10 出版日期:2013-04-01 发布日期:2013-04-23
  • 通讯作者: 王科俊
  • 作者简介:王科俊(1962-),男,吉林吉林人,教授,博士生导师,博士,主要研究方向:模式识别、神经网络;吕卓纹(1988-),女,吉林四平人,博士研究生,主要研究方向:模式识别、图像处理;孙国振(1985-),男,山东滕州人,硕士研究生,主要研究方向:视频监控;阎涛(1984-),男,河北衡水人,博士研究生,主要研究方向:模式识别、图像处理。
  • 基金资助:

    国家863计划项目(2008AA01Z148);国家自然科学基金资助项目(71173177);博士点专项科研基金资助项目(20102304110004)

Human behavior recognition based on stratified fractal conditional random field

WANG Kejun,LV Zhuowen,SUN Guozhen,YAN Tao   

  1. College of Automation, Harbin Engineering University, Harbin Heilongjiang 150001, China
  • Received:2012-10-16 Revised:2012-12-10 Online:2013-04-01 Published:2013-04-23
  • Contact: WANG Kejun

摘要: 针对隐条件随机场(HCRF)的实时性问题和隐动态条件随机场(LDCRF)行为转换时的标记偏差问题,提出了一种基于分层分数条件随机场(SFCRF)模型的行为识别算法。该算法改进了LDCRF,并提出分数标记的概念,将人体行为的完整性和有向性具体化。实验结果表明,该算法取得了比条件随机场(CRF)、HCRF和LDCRF更好的识别效果。

关键词: 隐条件随机场, 隐动态条件随机场, 标记偏差, 行为识别, 条件随机场

Abstract: In view of real-time issue of the Hidden Conditional Random Field (HCRF) and marked deviation problem of the Latent-Dynamic Conditional Random Field (LDCRF) during behavior transforming, this article proposed a kind of behavior recognition algorithm based on Stratified Fractal Conditional Random Field (SFCRF). The proposed algorithm improved LDCRF and put forward the concept of score mark, which made the integrity and direction of human behavior specific. The experimental results show that the proposed algorithm can obtain better recognition effect than Conditional Random Field (CRF), HCRF and LDCRF.

Key words: Hidden Conditional Random Field (HCRF), Latent-Dynamic Conditional Random Field (LDCRF), marked deviation, behavior recognition, Conditional Random Field (CRF)

中图分类号: