Head and upper body pose estimation in team sport videos

Masaki Hayashi, Taiki Yamamoto, Yoshimitsu Aoki, Kyoko Ohshima, Masamoto Tanabiki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

We propose a head and upper body pose estimation method in low-resolution team sports videos such as for American Football or Hockey, where all players wear helmets and often lean forward. Compared to the pedestrian cases in surveillance videos, head pose estimation technique for team sports videos has to deal with various types of activities (poses) and image scales according to the position of the player in the field. Using both the pelvis aligned player tracker and the head tracker, our system tracks the player's pelvis and head positions, which results in estimation of player's 2D spine. Then, we estimate the head and upper body orientations independently with random decision forest classifiers learned from a dataset including multiple-scale images. Integrating upper body direction and 2D spine pose, we also estimate the 3D spine pose of the player. Experiments show our method can estimate head and upper body pose accurately for sports players with intensive movement even without any temporal filtering techniques by focusing on the upper body region.

Original languageEnglish
Title of host publicationProceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013
PublisherIEEE Computer Society
Pages754-759
Number of pages6
DOIs
Publication statusPublished - 2013
Event2013 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013 - Naha, Okinawa, Japan
Duration: 2013 Nov 52013 Nov 8

Other

Other2013 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013
CountryJapan
CityNaha, Okinawa
Period13/11/513/11/8

Fingerprint

Sports
Classifiers
Wear of materials
Experiments

Keywords

  • Body pose estimation
  • Head pose estimation
  • Head tracking
  • Team sports analysis

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Hayashi, M., Yamamoto, T., Aoki, Y., Ohshima, K., & Tanabiki, M. (2013). Head and upper body pose estimation in team sport videos. In Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013 (pp. 754-759). [6778425] IEEE Computer Society. https://doi.org/10.1109/ACPR.2013.177

Head and upper body pose estimation in team sport videos. / Hayashi, Masaki; Yamamoto, Taiki; Aoki, Yoshimitsu; Ohshima, Kyoko; Tanabiki, Masamoto.

Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013. IEEE Computer Society, 2013. p. 754-759 6778425.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hayashi, M, Yamamoto, T, Aoki, Y, Ohshima, K & Tanabiki, M 2013, Head and upper body pose estimation in team sport videos. in Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013., 6778425, IEEE Computer Society, pp. 754-759, 2013 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013, Naha, Okinawa, Japan, 13/11/5. https://doi.org/10.1109/ACPR.2013.177
Hayashi M, Yamamoto T, Aoki Y, Ohshima K, Tanabiki M. Head and upper body pose estimation in team sport videos. In Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013. IEEE Computer Society. 2013. p. 754-759. 6778425 https://doi.org/10.1109/ACPR.2013.177
Hayashi, Masaki ; Yamamoto, Taiki ; Aoki, Yoshimitsu ; Ohshima, Kyoko ; Tanabiki, Masamoto. / Head and upper body pose estimation in team sport videos. Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013. IEEE Computer Society, 2013. pp. 754-759
@inproceedings{164c129346ce4cefb28247af0a7d8b32,
title = "Head and upper body pose estimation in team sport videos",
abstract = "We propose a head and upper body pose estimation method in low-resolution team sports videos such as for American Football or Hockey, where all players wear helmets and often lean forward. Compared to the pedestrian cases in surveillance videos, head pose estimation technique for team sports videos has to deal with various types of activities (poses) and image scales according to the position of the player in the field. Using both the pelvis aligned player tracker and the head tracker, our system tracks the player's pelvis and head positions, which results in estimation of player's 2D spine. Then, we estimate the head and upper body orientations independently with random decision forest classifiers learned from a dataset including multiple-scale images. Integrating upper body direction and 2D spine pose, we also estimate the 3D spine pose of the player. Experiments show our method can estimate head and upper body pose accurately for sports players with intensive movement even without any temporal filtering techniques by focusing on the upper body region.",
keywords = "Body pose estimation, Head pose estimation, Head tracking, Team sports analysis",
author = "Masaki Hayashi and Taiki Yamamoto and Yoshimitsu Aoki and Kyoko Ohshima and Masamoto Tanabiki",
year = "2013",
doi = "10.1109/ACPR.2013.177",
language = "English",
pages = "754--759",
booktitle = "Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Head and upper body pose estimation in team sport videos

AU - Hayashi, Masaki

AU - Yamamoto, Taiki

AU - Aoki, Yoshimitsu

AU - Ohshima, Kyoko

AU - Tanabiki, Masamoto

PY - 2013

Y1 - 2013

N2 - We propose a head and upper body pose estimation method in low-resolution team sports videos such as for American Football or Hockey, where all players wear helmets and often lean forward. Compared to the pedestrian cases in surveillance videos, head pose estimation technique for team sports videos has to deal with various types of activities (poses) and image scales according to the position of the player in the field. Using both the pelvis aligned player tracker and the head tracker, our system tracks the player's pelvis and head positions, which results in estimation of player's 2D spine. Then, we estimate the head and upper body orientations independently with random decision forest classifiers learned from a dataset including multiple-scale images. Integrating upper body direction and 2D spine pose, we also estimate the 3D spine pose of the player. Experiments show our method can estimate head and upper body pose accurately for sports players with intensive movement even without any temporal filtering techniques by focusing on the upper body region.

AB - We propose a head and upper body pose estimation method in low-resolution team sports videos such as for American Football or Hockey, where all players wear helmets and often lean forward. Compared to the pedestrian cases in surveillance videos, head pose estimation technique for team sports videos has to deal with various types of activities (poses) and image scales according to the position of the player in the field. Using both the pelvis aligned player tracker and the head tracker, our system tracks the player's pelvis and head positions, which results in estimation of player's 2D spine. Then, we estimate the head and upper body orientations independently with random decision forest classifiers learned from a dataset including multiple-scale images. Integrating upper body direction and 2D spine pose, we also estimate the 3D spine pose of the player. Experiments show our method can estimate head and upper body pose accurately for sports players with intensive movement even without any temporal filtering techniques by focusing on the upper body region.

KW - Body pose estimation

KW - Head pose estimation

KW - Head tracking

KW - Team sports analysis

UR - http://www.scopus.com/inward/record.url?scp=84899126123&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84899126123&partnerID=8YFLogxK

U2 - 10.1109/ACPR.2013.177

DO - 10.1109/ACPR.2013.177

M3 - Conference contribution

SP - 754

EP - 759

BT - Proceedings - 2nd IAPR Asian Conference on Pattern Recognition, ACPR 2013

PB - IEEE Computer Society

ER -