[1]范晓杰,宣士斌,唐 凤.基于混合时空特征描述子的人体动作识别[J].计算机技术与发展,2018,28(02):98-101.[doi:10．3969/j．issn．1673－629X．2018．02．022]
　FAN Xiao-jie,XUAN Shi-bin,TANG Feng.Realistic Human Action Recognition Based on Mixed Spatio-temporal Feature Descriptor[J].,2018,28(02):98-101.[doi:10．3969/j．issn．1673－629X．2018．02．022]
点击复制

基于混合时空特征描述子的人体动作识别()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
28
期数:
2018年02期
页码:
98-101
栏目:
智能、算法、系统工程
出版日期:
2018-02-10

文章信息/Info

Title:
Realistic Human Action Recognition Based on Mixed Spatio-temporal Feature Descriptor
文章编号:
1673－629X(2018)02－0098－04
作者:
范晓杰; 宣士斌; 唐 凤
广西民族大学 信息科学与工程学院，广西 南宁 530006
Author(s):
FAN Xiao-jie; XUAN Shi-bin; TANG Feng
School of Information Science and Engineering，Guangxi University for Nationlities，Nanning 530006，China
关键词:
时空兴趣点; 3D 有向直方图; 光流直方图; 自组织特征映射
Keywords:
spatio－temporal interest points (STIPs); orientation histograms of 3D gradient orientations (3DHOG); optical flow histogram(HOF); self－organizing feature map (SOM)
分类号:
TP31
DOI:
10．3969/j．issn．1673－629X．2018．02．022
文献标志码:
A
摘要:
 针对基于局部时空特征的行为识别中获取高效兴趣点、合理描述兴趣点及表征运动特征等关键问题，提出一种基于混合时空特征和 SOM 网络的新的行为识别框架。首先，从输入视频中提取出多尺度的 Dollar 时空兴趣点，并由时空兴趣点提取用于描述局部运动区域的视频块。然后，提出多向投影的光流直方图(DPHOF)构造方法，并与 3D 梯度方向直方图(HOG3D)结合描述视频块;利用 SOM 构造全局视频描述子。最后，用 K 最近邻(KNN)进行分类。对该方法在 KTH 和 UCF－YT 数据集上进行了验证，取得了很好的识别效果。实验结果表明，提出的 DPHOF 描述符能高效表示时空兴趣点，并优于HOG3D 和 HOF 的描述性，且由 SOM 构造出的全局视频描述子可以高效地表示视频特征，该方法具有更好的识别结果。
 
Abstract:
In view of key problems like efficient obtainment and reasonable description of interest points，and characterization of movement in human action recognition based on local features of time and space，we present a new action recognition framework based on mixed spacetime feature and SOM．Firstly，the multi－scale Dollar’s spatio－temporal interest points are extracted from the input video，and then the video block of describing local motion region is extracted by means of spatio－temporal interest points．Furthermore，we propose a novel multidirec-
tional projection optical flow histogram (DPHOF) descriptor to represent the video volume combined with the orientation histograms of 3D gradient orientations (3DHOG) and use SOM to generate the global video descriptor．Finally，the KNN is employed as classifier．This method is validated on the KTH and UCF－YT datasets with good recognition results．Experiment shows that the DPHOF descriptor proposed can efficiently represent the spatio－temporal interest points and is better than HOG3D and HOF．And the global video descriptor constructed by SOM can express the video features efficiently．The proposed method has better recognition effect．

相似文献/References:

[1]王博,李燕.视频序列中的时空兴趣点检测及其自适应分析[J].计算机技术与发展,2014,24(04):49.
　WANG Bo,LI Yan.Space-time Interest Points Detection in Video Sequence and Its Adaptive Analysis[J].,2014,24(02):49.
[2]金壮壮,曹江涛,姬晓飞.多源信息融合的双人交互行为识别算法研究[J].计算机技术与发展,2018,28(10):32.[doi:10.3969/ j. issn.1673-629X.2018.10.007]
　JIN Zhuang-zhuang,CAO Jiang-tao,JI Xiao-fei.Research on Human Interaction Recognition Algorithm Based on Multi-source Information Fusion[J].,2018,28(02):32.[doi:10.3969/ j. issn.1673-629X.2018.10.007]

更新日期/Last Update: 2018-03-29