南京大学学报(自然科学版) ›› 2012, Vol. 48 ›› Issue (1): 7783.
杜吉祥**,郭一兰,翟传敏
Du Ji一Xiang ,Guo Yi一Lan,Zhai Chuan-Ming
摘要: 提出一种基于局部时空兴趣点特征包的电影中复杂事件检索与识别的为一法.该为一法先将一个独立的事件视频序列表示成一个局部时空兴趣点特征包,再将此特征包与支持向量机相结合用于识别
事件.该为一法使用局部时空特征描述子来捕捉视频中的局部事件,可以适应事件的模式的不同的大小和速度.为了验证该为一法的有效性,使用了Hollywood视频数据库,其中的镜头序列收集自32部不同的
Hollywood电影,包含了8个事件类别.和其他相关的为一法相比,实验结果证明本文提出的为一法明显提高了平均正确率和平均查准率.
[1]Li Z D, Fei X L. Research on the concept-based information retrieval model. Journal of Nanjing University(Natural Sciences),2002,38(1): 99-109.(李振东,费翔林.基于概念的信息检索模型研究.南京大学学报(自然科学),2002,38(1):99一109). [2]Wu P, Hsieh J W, Chcng J C,et al. Human smoking event detection using visual interaction clues. International Conference on Pattern Rer ognition, 2010,4344一4347. [3]Laptcv I,Lindeberg T.Spacrtime interest points, IEEE international Conference on Com- puter Vision, 2003,132一139. [4]Dollar P,Rahaud V,Cottrell U, et al. Behav- for recognition via sparse spatio-temporal fea- tures. Joint IEEE international Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance,2005,65一72. [5]Jhuang H,Serre T,Wolf I.,et al. A biologi cally inspired system for action recognition. IEEE international Conference on Computer Vi sion, 2007,1一8. [6]Oikonomopoulos A,Patras I,Pantic M. Spatio- temporal salient points for visual recognition of human actions, IEEE Transactions on Systems Man and Cybernetics, Part B; Cybernetics 2006,36(3):710一719. [7]Willems G, T uytelaars T,VanUool L. An effi cient dense and scale-invariant spatio-temporal interest point detector. European Conference on Computer Vision, 2008,650一663. [8]Wong S F, Cipolla R. Extracting spatio-tempo ral interest points using global Information. IEEE international Conference on Computer Vi sion, 2007,1一8. [9]Klaser A,Marszalek M, Schmid C. A spatio- temporal descriptor based on 3D-gradients.British Machine Vision Conference,2008, 995一1004. [10]Laptev I,Lindeberg T.Local descriptors for spatio-temporal recognition. Lecture Notes in Computer Science. Springer-Verlag, 2004,3667.91一103. [11]Laptev I,Marszalck M,Schmid C,et al. Learning realistic human actions from IEEE Conference on Computer Vision and Pat-tern Recognition,2008,1~8 [12]Scovanner P,Ali S, Shah M. A 3-dimensional SIFT descriptor and its application to action rer ognition. ACM international Conference on Multimedia, 2007,357一360. [13]Laptev I,Patrick P. Retrieving actions in mov- ies. IEEE international Conference on Computer Vision, 2007,1一8. [14]Bian Z Q, Zhang X G. Pattern Recognition. The 2nd Edition. Beijing; Tsinghua University Press, 2000, 280-283.(边肇祺,张学工.模式识别.第二版.北京:清华大学出版社,2000, 280~283). [15]Schuldt C, Laptev l,Caputo B. Recognizing human actions:a local SVM approach, Interna tional Conference on Pattern Recognition, 2004,32~36 [16]Marszalek M, Laptev l,Schmid C. Actions in context, IEEE Conference on Computer Vision and Pattern Recognition, 2009,2929一2936. [17]Heng W, Muhammad M U, Alexander K,et al. Evaluation of local spatio-temporal features for action recognition. British Machine Vision Conference,2009,127一138. movies. |
No related articles found! |
|