基于区域形状与运动特征的实时行为识别

计算机科学 ›› 2013, Vol. 40 ›› Issue (5): 261-265.

基于区域形状与运动特征的实时行为识别

裴利沈,董乐,赵雪专,任鹏

电子科技大学计算机科学与工程学院成都611731;电子科技大学计算机科学与工程学院成都611731;中国科学院成都计算机应用研究所成都610041;电子科技大学空天科学技术研究院成都611731

出版日期:2018-11-16 发布日期:2018-11-16
基金资助:
本文受国家自然科学基金重大研究计划(91024026),国家自然科学基金(61003123,61105005),中央高校基本科研业务费(ZYGX2011X014,9)专项资金资助

Real-time Action Recognition Based on Zone Shapes and Motion Features

PEI Li-shen,DONG Le,ZHAO Xue-zhuan and REN Peng

Online:2018-11-16 Published:2018-11-16

摘要/Abstract

摘要： 提出了一种基于推广的Hu不变矩特征的实时行为识别方法。首先,对Hu不变矩进行改进,使其在离散情况下同时具有平移、旋转和比例不变性。然后,结合运动目标的速度将目标行为刻画成结合Hu矩新特征和速度特征的13维特性向量。其中,Hu矩新特征表征了行为的区域形状特性,速度特征反映了行为的运动特性。随后采用预先定义的一些行为作为先验知识样本训练支持向量机,并最后使用其对待检测行为进行分类以达到行为识别的效果。所提方法计算效率高,能够实时检测人体行为。在处理实拍视频数据的实验中,该方法表现出了理想的处理效率以及识别精度。

关键词: 行为识别,区域形状,Hu不变矩,运动特征

Abstract: This paper presented an efficient action recognition method based on Hu moment invariant features．Firstly,the Hu moment invariants were refined to be new features that are translation,rotation and scale invariant．Then an action was characterized by a 13-dimensional feature vector consisting of both Hu moment features and action speed features．The Hu moment features represent the Zone shape of the action,and the action speed features exhibit certain motion characteristics．Finally,a support vector machine(SVM),which is trained using labeled action frames,was applied to classify test sample actions into different categories．The proposed method is performed on real-world videos and achieves acceptable recognition rates with desirable computational efficiencies.

Key words: Behavior recognition,Zone shape,Hu moment invariant,Motion feature

裴利沈,董乐,赵雪专,任鹏. 基于区域形状与运动特征的实时行为识别[J]. 计算机科学, 2013, 40(5): 261-265. https://doi.org/

PEI Li-shen,DONG Le,ZHAO Xue-zhuan and REN Peng. Real-time Action Recognition Based on Zone Shapes and Motion Features[J]. Computer Science, 2013, 40(5): 261-265. https://doi.org/

参考文献

[1] Bobick A F,Davis J W．The recognition of human movement using temporal templates[C]∥IEEE International Conference on Computer Vision．2009
[2] Shechtman E,Irani M．Space-time behavior based correlation[C]∥IEEE Conference on Computer Vision and Pattern Recognition．2009
[3] Rodriguez M D,Ahmed J,Shah M．Action mach a spatio-temporal maximum average correlation height filter for action recognition[C]∥IEEE Conference on Computer Vision and Pattern Recognition．2008
[4] Hu Yu-xiao,Cao Liang-liang,Lv Feng-jun,et al.Action detection in complex scenes with spatial and temporal ambiguities[C]∥IEEE International Conference on Computer Vison．2009
[5] Hu M K．Visual pattern recognition by moment invariants[C]∥IRE Transaction Information theory．1962:179-182
[6] Wong Y R．Scene matching with invariant moments[J]．Computer Graphics and Image Processing,1978,8:16-24
[7] 丁明跃,常金玲,彭嘉雄．不变矩算法研究[J]．数据采集与处理,1992,7(1):1-9
[8] 吕洪涛,周继成．离散状态下的不变矩算法研究[J]．数据采集与处理,1993,8(2):151-155
[9] 刘进,张天序．图像不变矩的推广[J]．计算机学报,2004,7(5):668-674
[10] 印勇,张毅,刘丹平．基于改进Hu矩的异常行为识别[J]．计算机技术与发展,2009,9(9):90-92
[11] Zhang Guang-yuan,Cheng Bo,Feng Rui-jia,et al．Real-time drivereye detection method using Support Vector Machine with Hu invariant moments[J]．IEEE International Conference on Machine Learning and Cybernetics,2008(5):2999-3004
[12] Hwang C-L,Lee H-W．The command control by hand gesturewith Hu and contour sequence moments and probability neural network[C]∥IEEE International Conference on Systems,Man,and Cybernetics．2011:2056-2061
[13] Mu Hong-bo,Qi Da-wei．Pattern Recognition of Wood Defects Types Based on Hu Invariant Moments[C]∥2^nd International Congress on Image and Signal Processing．2009:1-5
[14] Safari M,Harandi M T.A SVM-based Method for Face Recognition Using a Wavelet-PCA Representation of Faces[C]∥IEEE International Conference on Image Processing．2004:853-856
[15] Schuldt C,Laptev I．Recognizing Human Actions:A local SVM Approach[C]∥International Conference on Pattern Recognition．2004

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed