Multi-agent event recognition by preservation of spatiotemporal relationships between probabilistic models
作者：
Highlights：
• Modeling/recognition of multi-agent activities (American football plays).
• Activities modeled as graphs, inexact graph matching used for comparison.
• Single-agent activity represented as motion patterns, modeled as graph nodes.
• Spatio-temporal relationships between single-agent behaviors modeled as graph edges.
• We present our own dataset of football plays called “UCF Football”.
摘要
•Modeling/recognition of multi-agent activities (American football plays).•Activities modeled as graphs, inexact graph matching used for comparison.•Single-agent activity represented as motion patterns, modeled as graph nodes.•Spatio-temporal relationships between single-agent behaviors modeled as graph edges.•We present our own dataset of football plays called “UCF Football”.
论文关键词：Football play recognition,Multi-agent activity modeling and recognition,Graph matching,Lie algebra
论文评审过程：Received 12 November 2012, Revised 22 May 2013, Accepted 4 June 2013, Available online 11 June 2013.
论文官网地址：https://doi.org/10.1016/j.imavis.2013.06.004
