Gesture spotting for low-resolution sports video annotation

Myung Cheol Roh, Bill Christmas, Joseph Kittler, Seong Whan Lee

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Human gesture recognition plays an important role in automating the analysis of video material at a high level. Especially in sports videos, the determination of the player's gestures is a key task. In many sports views, the camera covers a large part of the sports arena, resulting in low resolution of the player's region. Moreover, the camera is not static, but moves dynamically around its optical center, i.e. pan/tilt/zoom camera. These factors make the determination of the player's gestures a challenging task. To overcome these problems, we propose a posture descriptor that is robust to shape corruption of the player's silhouette, and a gesture spotting method that is robust to noisy sequences of data and needs only a small amount of training data. The proposed posture descriptor extracts the feature points of a shape, based on the curvature scale space (CSS) method. The use of CSS makes this method robust to local noise, and our method is also robust to significant shape corruption of the player's silhouette. The proposed spotting method provides probabilistic similarity and is robust to noisy sequences of data. It needs only a small number of training data sets, which is a very useful characteristic when it is difficult to obtain enough data for model training. In this paper, we conducted experiments spotting serve gestures using broadcast tennis play video. From our experiments, for 63 shots of playing tennis, some of which include a serve gesture and while some do not, it achieved 97.5% precision rate and 86.7% recall rate.

Original languageEnglish
Pages (from-to)1124-1137
Number of pages14
JournalPattern Recognition
Volume41
Issue number3
DOIs
Publication statusPublished - 2008 Mar 1

Fingerprint

Sports
Cameras
Gesture recognition
Experiments

Keywords

  • Gesture spotting
  • Low resolution video annotation
  • Posture descriptor
  • Posture determination

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Gesture spotting for low-resolution sports video annotation. / Roh, Myung Cheol; Christmas, Bill; Kittler, Joseph; Lee, Seong Whan.

In: Pattern Recognition, Vol. 41, No. 3, 01.03.2008, p. 1124-1137.

Research output: Contribution to journalArticle

Roh, Myung Cheol ; Christmas, Bill ; Kittler, Joseph ; Lee, Seong Whan. / Gesture spotting for low-resolution sports video annotation. In: Pattern Recognition. 2008 ; Vol. 41, No. 3. pp. 1124-1137.
@article{152734a2e0f74487b544b3946f3f8b17,
title = "Gesture spotting for low-resolution sports video annotation",
abstract = "Human gesture recognition plays an important role in automating the analysis of video material at a high level. Especially in sports videos, the determination of the player's gestures is a key task. In many sports views, the camera covers a large part of the sports arena, resulting in low resolution of the player's region. Moreover, the camera is not static, but moves dynamically around its optical center, i.e. pan/tilt/zoom camera. These factors make the determination of the player's gestures a challenging task. To overcome these problems, we propose a posture descriptor that is robust to shape corruption of the player's silhouette, and a gesture spotting method that is robust to noisy sequences of data and needs only a small amount of training data. The proposed posture descriptor extracts the feature points of a shape, based on the curvature scale space (CSS) method. The use of CSS makes this method robust to local noise, and our method is also robust to significant shape corruption of the player's silhouette. The proposed spotting method provides probabilistic similarity and is robust to noisy sequences of data. It needs only a small number of training data sets, which is a very useful characteristic when it is difficult to obtain enough data for model training. In this paper, we conducted experiments spotting serve gestures using broadcast tennis play video. From our experiments, for 63 shots of playing tennis, some of which include a serve gesture and while some do not, it achieved 97.5{\%} precision rate and 86.7{\%} recall rate.",
keywords = "Gesture spotting, Low resolution video annotation, Posture descriptor, Posture determination",
author = "Roh, {Myung Cheol} and Bill Christmas and Joseph Kittler and Lee, {Seong Whan}",
year = "2008",
month = "3",
day = "1",
doi = "10.1016/j.patcog.2007.07.013",
language = "English",
volume = "41",
pages = "1124--1137",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "3",

}

TY - JOUR

T1 - Gesture spotting for low-resolution sports video annotation

AU - Roh, Myung Cheol

AU - Christmas, Bill

AU - Kittler, Joseph

AU - Lee, Seong Whan

PY - 2008/3/1

Y1 - 2008/3/1

N2 - Human gesture recognition plays an important role in automating the analysis of video material at a high level. Especially in sports videos, the determination of the player's gestures is a key task. In many sports views, the camera covers a large part of the sports arena, resulting in low resolution of the player's region. Moreover, the camera is not static, but moves dynamically around its optical center, i.e. pan/tilt/zoom camera. These factors make the determination of the player's gestures a challenging task. To overcome these problems, we propose a posture descriptor that is robust to shape corruption of the player's silhouette, and a gesture spotting method that is robust to noisy sequences of data and needs only a small amount of training data. The proposed posture descriptor extracts the feature points of a shape, based on the curvature scale space (CSS) method. The use of CSS makes this method robust to local noise, and our method is also robust to significant shape corruption of the player's silhouette. The proposed spotting method provides probabilistic similarity and is robust to noisy sequences of data. It needs only a small number of training data sets, which is a very useful characteristic when it is difficult to obtain enough data for model training. In this paper, we conducted experiments spotting serve gestures using broadcast tennis play video. From our experiments, for 63 shots of playing tennis, some of which include a serve gesture and while some do not, it achieved 97.5% precision rate and 86.7% recall rate.

AB - Human gesture recognition plays an important role in automating the analysis of video material at a high level. Especially in sports videos, the determination of the player's gestures is a key task. In many sports views, the camera covers a large part of the sports arena, resulting in low resolution of the player's region. Moreover, the camera is not static, but moves dynamically around its optical center, i.e. pan/tilt/zoom camera. These factors make the determination of the player's gestures a challenging task. To overcome these problems, we propose a posture descriptor that is robust to shape corruption of the player's silhouette, and a gesture spotting method that is robust to noisy sequences of data and needs only a small amount of training data. The proposed posture descriptor extracts the feature points of a shape, based on the curvature scale space (CSS) method. The use of CSS makes this method robust to local noise, and our method is also robust to significant shape corruption of the player's silhouette. The proposed spotting method provides probabilistic similarity and is robust to noisy sequences of data. It needs only a small number of training data sets, which is a very useful characteristic when it is difficult to obtain enough data for model training. In this paper, we conducted experiments spotting serve gestures using broadcast tennis play video. From our experiments, for 63 shots of playing tennis, some of which include a serve gesture and while some do not, it achieved 97.5% precision rate and 86.7% recall rate.

KW - Gesture spotting

KW - Low resolution video annotation

KW - Posture descriptor

KW - Posture determination

UR - http://www.scopus.com/inward/record.url?scp=35448984571&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35448984571&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2007.07.013

DO - 10.1016/j.patcog.2007.07.013

M3 - Article

AN - SCOPUS:35448984571

VL - 41

SP - 1124

EP - 1137

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 3

ER -