Facial signatures for fast individual retrieval from video dataset

Pengyi Hao, Seiichiro Kamata

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

1 Citation (Scopus)

Abstract

The topic of retrieving videos containing a desired person by using the content of faces without any help of textual information has many interesting applications like video surveillance, social network, video mining, etc. However, face-by-face matching leads to an unacceptable response time for a video dataset with a large number of detected faces and may also reduce the accuracy of searching. Therefore, in this paper we propose a scheme to generate facial signatures for fast retrieving videos containing the same person with a query. First, we summarize each video as a set of person-oriented individuals based on detected faces, which are represented as high dimensional vectors in a feature space. Then, each person with a collection of high dimensional vectors is projected to a compact and reduced dimensionality representation that is called facial signature for this person. The projection is realized by constructing a matcher using linear discriminant analysis with maximum correntropy criterion optimization. In this research, two kinds of signatures are provided, which are called 1D facial signature and 2D facial signature. The proposed searching scheme can support two types of queries: face image and video clip. Evaluations on a large dataset of videos show reliable measurement of similarities by using facial signature to represent each person generated from videos and also demonstrate that the proposed searching scheme has the potential to substantially reduce the response time and slightly increase the mean average precision of retrieval.

Original languageEnglish
Title of host publicationProceedings - IEEE International Conference on Multimedia and Expo
DOIs
Publication statusPublished - 2013
Event2013 IEEE International Conference on Multimedia and Expo, ICME 2013 - San Jose, CA
Duration: 2013 Jul 152013 Jul 19

Other

Other2013 IEEE International Conference on Multimedia and Expo, ICME 2013
CitySan Jose, CA
Period13/7/1513/7/19

Fingerprint

Discriminant analysis

Keywords

  • Face retrieval
  • Facial signature
  • Indexing

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications

Cite this

Hao, P., & Kamata, S. (2013). Facial signatures for fast individual retrieval from video dataset. In Proceedings - IEEE International Conference on Multimedia and Expo [6607450] https://doi.org/10.1109/ICME.2013.6607450

Facial signatures for fast individual retrieval from video dataset. / Hao, Pengyi; Kamata, Seiichiro.

Proceedings - IEEE International Conference on Multimedia and Expo. 2013. 6607450.

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

Hao, P & Kamata, S 2013, Facial signatures for fast individual retrieval from video dataset. in Proceedings - IEEE International Conference on Multimedia and Expo., 6607450, 2013 IEEE International Conference on Multimedia and Expo, ICME 2013, San Jose, CA, 13/7/15. https://doi.org/10.1109/ICME.2013.6607450
Hao P, Kamata S. Facial signatures for fast individual retrieval from video dataset. In Proceedings - IEEE International Conference on Multimedia and Expo. 2013. 6607450 https://doi.org/10.1109/ICME.2013.6607450
Hao, Pengyi ; Kamata, Seiichiro. / Facial signatures for fast individual retrieval from video dataset. Proceedings - IEEE International Conference on Multimedia and Expo. 2013.
@inproceedings{9eab35565c814298aa012100514c58e7,
title = "Facial signatures for fast individual retrieval from video dataset",
abstract = "The topic of retrieving videos containing a desired person by using the content of faces without any help of textual information has many interesting applications like video surveillance, social network, video mining, etc. However, face-by-face matching leads to an unacceptable response time for a video dataset with a large number of detected faces and may also reduce the accuracy of searching. Therefore, in this paper we propose a scheme to generate facial signatures for fast retrieving videos containing the same person with a query. First, we summarize each video as a set of person-oriented individuals based on detected faces, which are represented as high dimensional vectors in a feature space. Then, each person with a collection of high dimensional vectors is projected to a compact and reduced dimensionality representation that is called facial signature for this person. The projection is realized by constructing a matcher using linear discriminant analysis with maximum correntropy criterion optimization. In this research, two kinds of signatures are provided, which are called 1D facial signature and 2D facial signature. The proposed searching scheme can support two types of queries: face image and video clip. Evaluations on a large dataset of videos show reliable measurement of similarities by using facial signature to represent each person generated from videos and also demonstrate that the proposed searching scheme has the potential to substantially reduce the response time and slightly increase the mean average precision of retrieval.",
keywords = "Face retrieval, Facial signature, Indexing",
author = "Pengyi Hao and Seiichiro Kamata",
year = "2013",
doi = "10.1109/ICME.2013.6607450",
language = "English",
isbn = "9781479900152",
booktitle = "Proceedings - IEEE International Conference on Multimedia and Expo",

}

TY - GEN

T1 - Facial signatures for fast individual retrieval from video dataset

AU - Hao, Pengyi

AU - Kamata, Seiichiro

PY - 2013

Y1 - 2013

N2 - The topic of retrieving videos containing a desired person by using the content of faces without any help of textual information has many interesting applications like video surveillance, social network, video mining, etc. However, face-by-face matching leads to an unacceptable response time for a video dataset with a large number of detected faces and may also reduce the accuracy of searching. Therefore, in this paper we propose a scheme to generate facial signatures for fast retrieving videos containing the same person with a query. First, we summarize each video as a set of person-oriented individuals based on detected faces, which are represented as high dimensional vectors in a feature space. Then, each person with a collection of high dimensional vectors is projected to a compact and reduced dimensionality representation that is called facial signature for this person. The projection is realized by constructing a matcher using linear discriminant analysis with maximum correntropy criterion optimization. In this research, two kinds of signatures are provided, which are called 1D facial signature and 2D facial signature. The proposed searching scheme can support two types of queries: face image and video clip. Evaluations on a large dataset of videos show reliable measurement of similarities by using facial signature to represent each person generated from videos and also demonstrate that the proposed searching scheme has the potential to substantially reduce the response time and slightly increase the mean average precision of retrieval.

AB - The topic of retrieving videos containing a desired person by using the content of faces without any help of textual information has many interesting applications like video surveillance, social network, video mining, etc. However, face-by-face matching leads to an unacceptable response time for a video dataset with a large number of detected faces and may also reduce the accuracy of searching. Therefore, in this paper we propose a scheme to generate facial signatures for fast retrieving videos containing the same person with a query. First, we summarize each video as a set of person-oriented individuals based on detected faces, which are represented as high dimensional vectors in a feature space. Then, each person with a collection of high dimensional vectors is projected to a compact and reduced dimensionality representation that is called facial signature for this person. The projection is realized by constructing a matcher using linear discriminant analysis with maximum correntropy criterion optimization. In this research, two kinds of signatures are provided, which are called 1D facial signature and 2D facial signature. The proposed searching scheme can support two types of queries: face image and video clip. Evaluations on a large dataset of videos show reliable measurement of similarities by using facial signature to represent each person generated from videos and also demonstrate that the proposed searching scheme has the potential to substantially reduce the response time and slightly increase the mean average precision of retrieval.

KW - Face retrieval

KW - Facial signature

KW - Indexing

UR - http://www.scopus.com/inward/record.url?scp=84885622735&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84885622735&partnerID=8YFLogxK

U2 - 10.1109/ICME.2013.6607450

DO - 10.1109/ICME.2013.6607450

M3 - Conference contribution

SN - 9781479900152

BT - Proceedings - IEEE International Conference on Multimedia and Expo

ER -