Audio-based event detection for sports video

M. Baillie, J.M. Jose

Research output: Contribution to journalArticle

27 Citations (Scopus)

Abstract

In this paper, we present an audio-based event detection approach shown to be effective when applied to sports broadcast data. The main benefit of this approach is the ability to recognise patterns that indicate high levels of crowd response which can be correlated to key events. By applying Hidden Markov Model-based classifiers, where the predefined content classes are parameterised using Mel-Frequency Cepstral Coefficients, we were able to eliminate the need for defining a heuristic set of rules to determine event detection, thus avoiding a two-class approach shown not to be suitable for this problem. Experimentation indicated that this is an effective method for classifying crowd response in football matches, thus providing a basis for automatic indexing and summarisation.
Original languageEnglish
JournalLecture Notes in Computer Science
Publication statusPublished - 2003

Fingerprint

Automatic indexing
Event Detection
Hidden Markov models
Sports
Classifiers
Summarization
Indexing
Experimentation
Broadcast
Markov Model
Eliminate
Classifier
Heuristics
Model-based
Coefficient
Class

Keywords

  • information retrieval
  • learning objects
  • sport
  • mel-frequency cepstral coefficients

Cite this

@article{b18398587d7c4161aeb17885e6785b44,
title = "Audio-based event detection for sports video",
abstract = "In this paper, we present an audio-based event detection approach shown to be effective when applied to sports broadcast data. The main benefit of this approach is the ability to recognise patterns that indicate high levels of crowd response which can be correlated to key events. By applying Hidden Markov Model-based classifiers, where the predefined content classes are parameterised using Mel-Frequency Cepstral Coefficients, we were able to eliminate the need for defining a heuristic set of rules to determine event detection, thus avoiding a two-class approach shown not to be suitable for this problem. Experimentation indicated that this is an effective method for classifying crowd response in football matches, thus providing a basis for automatic indexing and summarisation.",
keywords = "information retrieval, learning objects, sport, mel-frequency cepstral coefficients",
author = "M. Baillie and J.M. Jose",
year = "2003",
language = "English",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",

}

Audio-based event detection for sports video. / Baillie, M.; Jose, J.M.

In: Lecture Notes in Computer Science, 2003.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Audio-based event detection for sports video

AU - Baillie, M.

AU - Jose, J.M.

PY - 2003

Y1 - 2003

N2 - In this paper, we present an audio-based event detection approach shown to be effective when applied to sports broadcast data. The main benefit of this approach is the ability to recognise patterns that indicate high levels of crowd response which can be correlated to key events. By applying Hidden Markov Model-based classifiers, where the predefined content classes are parameterised using Mel-Frequency Cepstral Coefficients, we were able to eliminate the need for defining a heuristic set of rules to determine event detection, thus avoiding a two-class approach shown not to be suitable for this problem. Experimentation indicated that this is an effective method for classifying crowd response in football matches, thus providing a basis for automatic indexing and summarisation.

AB - In this paper, we present an audio-based event detection approach shown to be effective when applied to sports broadcast data. The main benefit of this approach is the ability to recognise patterns that indicate high levels of crowd response which can be correlated to key events. By applying Hidden Markov Model-based classifiers, where the predefined content classes are parameterised using Mel-Frequency Cepstral Coefficients, we were able to eliminate the need for defining a heuristic set of rules to determine event detection, thus avoiding a two-class approach shown not to be suitable for this problem. Experimentation indicated that this is an effective method for classifying crowd response in football matches, thus providing a basis for automatic indexing and summarisation.

KW - information retrieval

KW - learning objects

KW - sport

KW - mel-frequency cepstral coefficients

UR - http://www.springerlink.com/content/mwuhfqcaqwuk7n3p/

M3 - Article

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -