A topic model approach to represent and classify american football plays

Jagannadan Varadarajan, Indriyati Atmosukarto, Shaunak Ahuja, Bernard Ghanem, Narendra Ahuja

Research output: Contribution to conferencePaper

Abstract

We address the problem of modeling and classifying American Football offense teams' plays in video, a challenging example of group activity analysis. Automatic play classification will allow coaches to infer patterns and tendencies of opponents more efficiently, resulting in better strategy planning in a game. We define a football play as a unique combination of player trajectories. We develop a framework that uses player trajectories as inputs to MedLDA, a supervised topic model. The joint maximization of both likelihood and inter-class margins of MedLDA in learning the topics allows us to learn semantically meaningful play type templates, as well as, classify different play types with 70% average accuracy. Furthermore, this method is extended to analyze individual player roles in classifying each play type. We validate our method on a large dataset comprising 271 play clips from real-world football games, which will be made publicly available for future comparisons.

Original languageEnglish (US)
DOIs
StatePublished - Jan 1 2013
Event2013 24th British Machine Vision Conference, BMVC 2013 - Bristol, United Kingdom
Duration: Sep 9 2013Sep 13 2013

Other

Other2013 24th British Machine Vision Conference, BMVC 2013
CountryUnited Kingdom
CityBristol
Period9/9/139/13/13

Fingerprint

Trajectories
Planning

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Varadarajan, J., Atmosukarto, I., Ahuja, S., Ghanem, B., & Ahuja, N. (2013). A topic model approach to represent and classify american football plays. Paper presented at 2013 24th British Machine Vision Conference, BMVC 2013, Bristol, United Kingdom. https://doi.org/10.5244/C.27.64

A topic model approach to represent and classify american football plays. / Varadarajan, Jagannadan; Atmosukarto, Indriyati; Ahuja, Shaunak; Ghanem, Bernard; Ahuja, Narendra.

2013. Paper presented at 2013 24th British Machine Vision Conference, BMVC 2013, Bristol, United Kingdom.

Research output: Contribution to conferencePaper

Varadarajan, J, Atmosukarto, I, Ahuja, S, Ghanem, B & Ahuja, N 2013, 'A topic model approach to represent and classify american football plays', Paper presented at 2013 24th British Machine Vision Conference, BMVC 2013, Bristol, United Kingdom, 9/9/13 - 9/13/13. https://doi.org/10.5244/C.27.64
Varadarajan J, Atmosukarto I, Ahuja S, Ghanem B, Ahuja N. A topic model approach to represent and classify american football plays. 2013. Paper presented at 2013 24th British Machine Vision Conference, BMVC 2013, Bristol, United Kingdom. https://doi.org/10.5244/C.27.64
Varadarajan, Jagannadan ; Atmosukarto, Indriyati ; Ahuja, Shaunak ; Ghanem, Bernard ; Ahuja, Narendra. / A topic model approach to represent and classify american football plays. Paper presented at 2013 24th British Machine Vision Conference, BMVC 2013, Bristol, United Kingdom.
@conference{eb2dfbdef5a14e20b100b3934dcb6eb4,
title = "A topic model approach to represent and classify american football plays",
abstract = "We address the problem of modeling and classifying American Football offense teams' plays in video, a challenging example of group activity analysis. Automatic play classification will allow coaches to infer patterns and tendencies of opponents more efficiently, resulting in better strategy planning in a game. We define a football play as a unique combination of player trajectories. We develop a framework that uses player trajectories as inputs to MedLDA, a supervised topic model. The joint maximization of both likelihood and inter-class margins of MedLDA in learning the topics allows us to learn semantically meaningful play type templates, as well as, classify different play types with 70{\%} average accuracy. Furthermore, this method is extended to analyze individual player roles in classifying each play type. We validate our method on a large dataset comprising 271 play clips from real-world football games, which will be made publicly available for future comparisons.",
author = "Jagannadan Varadarajan and Indriyati Atmosukarto and Shaunak Ahuja and Bernard Ghanem and Narendra Ahuja",
year = "2013",
month = "1",
day = "1",
doi = "10.5244/C.27.64",
language = "English (US)",
note = "2013 24th British Machine Vision Conference, BMVC 2013 ; Conference date: 09-09-2013 Through 13-09-2013",

}

TY - CONF

T1 - A topic model approach to represent and classify american football plays

AU - Varadarajan, Jagannadan

AU - Atmosukarto, Indriyati

AU - Ahuja, Shaunak

AU - Ghanem, Bernard

AU - Ahuja, Narendra

PY - 2013/1/1

Y1 - 2013/1/1

N2 - We address the problem of modeling and classifying American Football offense teams' plays in video, a challenging example of group activity analysis. Automatic play classification will allow coaches to infer patterns and tendencies of opponents more efficiently, resulting in better strategy planning in a game. We define a football play as a unique combination of player trajectories. We develop a framework that uses player trajectories as inputs to MedLDA, a supervised topic model. The joint maximization of both likelihood and inter-class margins of MedLDA in learning the topics allows us to learn semantically meaningful play type templates, as well as, classify different play types with 70% average accuracy. Furthermore, this method is extended to analyze individual player roles in classifying each play type. We validate our method on a large dataset comprising 271 play clips from real-world football games, which will be made publicly available for future comparisons.

AB - We address the problem of modeling and classifying American Football offense teams' plays in video, a challenging example of group activity analysis. Automatic play classification will allow coaches to infer patterns and tendencies of opponents more efficiently, resulting in better strategy planning in a game. We define a football play as a unique combination of player trajectories. We develop a framework that uses player trajectories as inputs to MedLDA, a supervised topic model. The joint maximization of both likelihood and inter-class margins of MedLDA in learning the topics allows us to learn semantically meaningful play type templates, as well as, classify different play types with 70% average accuracy. Furthermore, this method is extended to analyze individual player roles in classifying each play type. We validate our method on a large dataset comprising 271 play clips from real-world football games, which will be made publicly available for future comparisons.

UR - http://www.scopus.com/inward/record.url?scp=84898460846&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898460846&partnerID=8YFLogxK

U2 - 10.5244/C.27.64

DO - 10.5244/C.27.64

M3 - Paper

AN - SCOPUS:84898460846

ER -