SEMANTICEVENT EXTRACTIONFROM BASKETBALLGAMES USINGMULTI-MODAL ANALYSIS

Yifan Zhang, Changsheng Xu, Jinqiao Wang, Hanqing Lu

Abstract

Inthis paper, wepresent anovel multi-modal framework forsemantic eventextraction frombasketball gamesbasedonwebcasting text andbroadcast video. We propose novel approaches to textanalysis forevent detection andsemantics extraction, video analysis forevent structure modeling andevent momentdetection, andtext/video alignment forevent boundary detection inthevideo. Compared withexisting approaches toeventdetection insports video whichrely heavily onlow-level features directly extracted fromvideo itself, ourapproach aimstobridge thesemantic gap between low-level features andhigh-level events andfacilitates personalization ofthesports video. Promising results arereported onreal-world video clips byusing text analysis, video analysis and text/video alignment.
BibTeX
@inproceedings{zhang2007semanticevent,
  author = {Yifan Zhang and Changsheng Xu and Jinqiao Wang and Hanqing Lu},
  title = {{SEMANTICEVENT EXTRACTIONFROM BASKETBALLGAMES USINGMULTI-MODAL ANALYSIS}},
  booktitle = {ICME},
  year = {2007},
  doi = {10.1109/icme.2007.4285119},
}

Loading related content...

Figure