期刊文献+

基于固定数采样的轮廓特征描述子方法

下载PDF
导出
摘要 本文在特征点的选取中使用固定数采样,降低特征点数目并且保存特征点的整体信息,然后采用质心边界距离描述子对轮廓进行描述,最后再将位置信息降维,生成一维的轮廓特征对目标运动状态进行描述。
作者 贾贺林
出处 《电子技术与软件工程》 2019年第7期62-62,共1页 ELECTRONIC TECHNOLOGY & SOFTWARE ENGINEERING
  • 相关文献

参考文献2

二级参考文献30

  • 1张德才.视频流中多人体跟踪算法研究[D].长春:吉林大学,2011.
  • 2Feng Z,Yang B, Li Y, et al. Real-time oriented behavior- driven 3D freehand tracking for direct interaction[J]. Pat- tern Recognition, 2013,46(2) : 590-608.
  • 3Poppe R. A survey on vision-based human action recogni- tion[J]. Image and vision computing, 20]0,28 (6) : 976- 990.
  • 4Gorelick L,Blank M,Shechtman E,et al. Actions as spac- e-time shapes[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2007,29( 12 ) : 2247-2253.
  • 5Niebles J C,Wang H, Fei-Fei L. Unsupervised learning of human action categories using spatial-temporal words [J]. International Journal of Computer Vision, 2008,79 (3) :299-318.
  • 6Derpanis K G, Sizintsev M, Cannons K J, et al. Action spotting and recognition based on a spatiotemporal orien- tation analysis[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(3) :527-540.
  • 7Zhou H, Wang L, Suter D. Human action recognition by feature-reduced Gaussian process classification[J]. Pat- tern Recognition Letters,2009,30(12):1059-1066,.
  • 8Babu R V, Ramakrishnan K R. Recognition of human ac-tions using motion history information extracted from the compressed video [J]. Image and Vision Computing, 2004,22(8) :597-607.
  • 9Yogameena B,Lakshmi S V, Archana M, et al. Human be havior classification using multi-class relevance vectoI machine[J]. Journal of Computer Science, 2010,6 (9) : 1021-1026.
  • 10He W, Yow K C, Guo Y. Recognition of human activities using a multiclass relevance vector machine[J]. Optical Engineering, 2012,51 (1) :017202;1-12.

共引文献20

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部