首页 | 本学科首页   官方微博 | 高级检索  
     

基于凸型模板的目标形态图像识别算法
引用本文:杜会斌,刘承禹.基于凸型模板的目标形态图像识别算法[J].计算机与网络,2011(14):49-52.
作者姓名:杜会斌  刘承禹
作者单位:[1]中国电子科技集团公司第五十四研究所,河北石家庄050081 [2]第二炮兵装备研究院,北京100085
摘    要:针对视频图像中目标形态的检测识别问题,提出了一种基于凸型模板修正算法对条形目标的检测、识别和定位技术。通过分析凸型模板识别条形目标的基本原理,提出引入倾斜角度因子的凸型模板改进算法,提高了条形目标识别的准确度,并以单人个体作为研究对象进行实验,结果表明这种方法运算复杂度低,对条形目标的识别率高,跟踪特性良好。

关 键 词:运动目标跟踪  凸型模板  肢体定位  姿态识别

Object Gesture Recognition Method Based on Torso Mask
DU Hui-bin,The,th Research Institute of CETC,Shijiazhuang Hebei,China,LIU Cheng-yu.Object Gesture Recognition Method Based on Torso Mask[J].China Computer & Network,2011(14):49-52.
Authors:DU Hui-bin  The  th Research Institute of CETC  Shijiazhuang Hebei  China  LIU Cheng-yu
Affiliation:Equipment Research Institute of Second Artillery Beijing 100085 China)
Abstract:The paper introduces a fast object tracking and behavior recognition technology applicable for human detection and gesture recognition in video and image processing filed. It analyzes the principle of torso mask and implementation steps of the improved method. This method increases the accuracy of torso location through introducing torso tilt angle factor. The experimental results show that the new method achieves lower complexity in calculation, higher recognition in object and better performance in tracking.
Keywords:object tracking  torso mask  body location  gesture recognition
本文献已被 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号