Clustering faces in movies using an automatically constructed social network

Mei Chen Yeh, Wen Po Wu

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Clustering faces in movies is a challenging task because faces in a feature-length film are relatively uncontrolled and vary widely in appearance. Such variations make it difficult to appropriately measure the similarity between faces under significantly different settings. In this article, the authors develop a method that improves face-clustering accuracy by incorporating the social context information inherent among characters in a movie. In particular, they study the relation of social network construction and face clustering and present a fusion scheme that eliminates ambiguities and bridges information from two fields. Experiments on real-world data show superior clustering performance compared with state-of-the-art methods. Furthermore, their method can help incrementally build a character's social network that is similar to a manually labeled example.

Original languageEnglish
Article number6818950
Pages (from-to)22-31
Number of pages10
JournalIEEE Multimedia
Volume21
Issue number2
DOIs
Publication statusPublished - 2014 Jan 1

Fingerprint

Fusion reactions
Experiments

Keywords

  • face clustering
  • movie content analysis
  • multimedia
  • social network

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Media Technology
  • Hardware and Architecture
  • Computer Science Applications

Cite this

Clustering faces in movies using an automatically constructed social network. / Yeh, Mei Chen; Wu, Wen Po.

In: IEEE Multimedia, Vol. 21, No. 2, 6818950, 01.01.2014, p. 22-31.

Research output: Contribution to journalArticle

@article{6838da452095427bb78371a9470b9a78,
title = "Clustering faces in movies using an automatically constructed social network",
abstract = "Clustering faces in movies is a challenging task because faces in a feature-length film are relatively uncontrolled and vary widely in appearance. Such variations make it difficult to appropriately measure the similarity between faces under significantly different settings. In this article, the authors develop a method that improves face-clustering accuracy by incorporating the social context information inherent among characters in a movie. In particular, they study the relation of social network construction and face clustering and present a fusion scheme that eliminates ambiguities and bridges information from two fields. Experiments on real-world data show superior clustering performance compared with state-of-the-art methods. Furthermore, their method can help incrementally build a character's social network that is similar to a manually labeled example.",
keywords = "face clustering, movie content analysis, multimedia, social network",
author = "Yeh, {Mei Chen} and Wu, {Wen Po}",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/MMUL.2014.24",
language = "English",
volume = "21",
pages = "22--31",
journal = "IEEE Multimedia",
issn = "1070-986X",
publisher = "IEEE Computer Society",
number = "2",

}

TY - JOUR

T1 - Clustering faces in movies using an automatically constructed social network

AU - Yeh, Mei Chen

AU - Wu, Wen Po

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Clustering faces in movies is a challenging task because faces in a feature-length film are relatively uncontrolled and vary widely in appearance. Such variations make it difficult to appropriately measure the similarity between faces under significantly different settings. In this article, the authors develop a method that improves face-clustering accuracy by incorporating the social context information inherent among characters in a movie. In particular, they study the relation of social network construction and face clustering and present a fusion scheme that eliminates ambiguities and bridges information from two fields. Experiments on real-world data show superior clustering performance compared with state-of-the-art methods. Furthermore, their method can help incrementally build a character's social network that is similar to a manually labeled example.

AB - Clustering faces in movies is a challenging task because faces in a feature-length film are relatively uncontrolled and vary widely in appearance. Such variations make it difficult to appropriately measure the similarity between faces under significantly different settings. In this article, the authors develop a method that improves face-clustering accuracy by incorporating the social context information inherent among characters in a movie. In particular, they study the relation of social network construction and face clustering and present a fusion scheme that eliminates ambiguities and bridges information from two fields. Experiments on real-world data show superior clustering performance compared with state-of-the-art methods. Furthermore, their method can help incrementally build a character's social network that is similar to a manually labeled example.

KW - face clustering

KW - movie content analysis

KW - multimedia

KW - social network

UR - http://www.scopus.com/inward/record.url?scp=84901436811&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84901436811&partnerID=8YFLogxK

U2 - 10.1109/MMUL.2014.24

DO - 10.1109/MMUL.2014.24

M3 - Article

AN - SCOPUS:84901436811

VL - 21

SP - 22

EP - 31

JO - IEEE Multimedia

JF - IEEE Multimedia

SN - 1070-986X

IS - 2

M1 - 6818950

ER -