一种基于简单特征量信息的快速行为识别算法  

A fast behavior recognition algorithm based on simple information of characteristic quantity

作　　者：纪彬[1] 张建平[1] 周大志[1] 李磊[1] 

机构地区：[1]黑龙江省科学院自动化研究所,哈尔滨150090

出　　处：《黑龙江科学》2016年第17期1-3,6,共4页Heilongjiang Science

摘　　要：人体行为识别技术是利用机器视觉,以人体的生物特征参数为基础,对人的数量、行为和身份等信息进行快速准确分析判断的一种实用技术。该技术可应用于安防监控、安全验证、军事、商业和工业等领域。提出一种基于简单特征量信息的快速行为识别算法,与当前主流的算法相比,有可用更少的训练数据和识别效果更加高效的特点。Human behavior recognition technology is a practical technique that makes rapid accurate analysis of the number of people,behavior,and identity information based on machine vision and body＇s biological characteristic parameter.This technology can be widely used in security monitoring,security,military,commerce,industry and many other areas.This paper proposes a recognition algorithm based on simple characteristic information fast behavior,it has characteristics of having fewer available training data and more efficient recognition effect compared with the current main trend algorithm.

关 键 词：特征量 行为识别 分类器 KTH数据库 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]