Nighttime face recognition at long distance: Cross-distance and cross-spectral matching

Hyunju Maeng, Shengcai Liao, Dongoh Kang, Seong Whan Lee, Anil K. Jain

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

23 Citations (Scopus)

Abstract

Automatic face recognition capability in surveillance systems is important for security applications. However, few studies have addressed the problem of outdoor face recognition at a long distance (over 100 meters) in both daytime and nighttime environments. In this paper, we first report on a system that we have designed to collect face image database at a long distance, called the Long Distance Heterogeneous Face Database (LDHF-DB) to advance research on this topic. The LDHF-DB contains face images collected in an outdoor environment at distances of 60 meters, 100 meters, and 150 meters, with both visible light (VIS) face images captured in daytime and near infrared (NIR) face images captured in nighttime. Given this database, we have conducted two types of cross-distance face matching (matching long-distance probe to 1-meter gallery) experiments: (i) intra-spectral (VIS to VIS) face matching, and (ii) cross-spectral (NIR to VIS) face matching. The proposed face recognition algorithm consists of following three major steps: (i) Gaussian filtering to remove high frequency noise, (ii) Scale Invariant Feature Transform (SIFT) in local image regions for feature representation, and (iii) a random subspace method to build discriminant subspaces for face recognition. Experimental results show that the proposed face recognition algorithm outperforms two commercial state-of-the-art face recognition SDKs (FaceVACS and PittPatt) for long distance face recognition in both daytime and nighttime operations. These results highlight the need for better data capture setup and robust face matching algorithms for cross spectral matching at distances greater than 100 meters.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages708-721
Number of pages14
Volume7725 LNCS
EditionPART 2
DOIs
Publication statusPublished - 2013 Apr 11
Event11th Asian Conference on Computer Vision, ACCV 2012 - Daejeon, Korea, Republic of
Duration: 2012 Nov 52012 Nov 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume7725 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other11th Asian Conference on Computer Vision, ACCV 2012
CountryKorea, Republic of
CityDaejeon
Period12/11/512/11/9

Fingerprint

Face recognition
Face Recognition
Face
Recognition Algorithm
Infrared radiation
Infrared
Subspace Methods
Data acquisition
Scale Invariant Feature Transform
Image Database
Mathematical transformations
Matching Algorithm
Discriminant
Surveillance
Probe
Filtering
Subspace
Experimental Results
Experiments

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Maeng, H., Liao, S., Kang, D., Lee, S. W., & Jain, A. K. (2013). Nighttime face recognition at long distance: Cross-distance and cross-spectral matching. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 2 ed., Vol. 7725 LNCS, pp. 708-721). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7725 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-37444-9_55

Nighttime face recognition at long distance : Cross-distance and cross-spectral matching. / Maeng, Hyunju; Liao, Shengcai; Kang, Dongoh; Lee, Seong Whan; Jain, Anil K.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7725 LNCS PART 2. ed. 2013. p. 708-721 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7725 LNCS, No. PART 2).

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

Maeng, H, Liao, S, Kang, D, Lee, SW & Jain, AK 2013, Nighttime face recognition at long distance: Cross-distance and cross-spectral matching. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 edn, vol. 7725 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 7725 LNCS, pp. 708-721, 11th Asian Conference on Computer Vision, ACCV 2012, Daejeon, Korea, Republic of, 12/11/5. https://doi.org/10.1007/978-3-642-37444-9_55
Maeng H, Liao S, Kang D, Lee SW, Jain AK. Nighttime face recognition at long distance: Cross-distance and cross-spectral matching. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 ed. Vol. 7725 LNCS. 2013. p. 708-721. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-642-37444-9_55
Maeng, Hyunju ; Liao, Shengcai ; Kang, Dongoh ; Lee, Seong Whan ; Jain, Anil K. / Nighttime face recognition at long distance : Cross-distance and cross-spectral matching. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7725 LNCS PART 2. ed. 2013. pp. 708-721 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).
@inproceedings{a2e845fcff094d53b88f0ca336d7fcc9,
title = "Nighttime face recognition at long distance: Cross-distance and cross-spectral matching",
abstract = "Automatic face recognition capability in surveillance systems is important for security applications. However, few studies have addressed the problem of outdoor face recognition at a long distance (over 100 meters) in both daytime and nighttime environments. In this paper, we first report on a system that we have designed to collect face image database at a long distance, called the Long Distance Heterogeneous Face Database (LDHF-DB) to advance research on this topic. The LDHF-DB contains face images collected in an outdoor environment at distances of 60 meters, 100 meters, and 150 meters, with both visible light (VIS) face images captured in daytime and near infrared (NIR) face images captured in nighttime. Given this database, we have conducted two types of cross-distance face matching (matching long-distance probe to 1-meter gallery) experiments: (i) intra-spectral (VIS to VIS) face matching, and (ii) cross-spectral (NIR to VIS) face matching. The proposed face recognition algorithm consists of following three major steps: (i) Gaussian filtering to remove high frequency noise, (ii) Scale Invariant Feature Transform (SIFT) in local image regions for feature representation, and (iii) a random subspace method to build discriminant subspaces for face recognition. Experimental results show that the proposed face recognition algorithm outperforms two commercial state-of-the-art face recognition SDKs (FaceVACS and PittPatt) for long distance face recognition in both daytime and nighttime operations. These results highlight the need for better data capture setup and robust face matching algorithms for cross spectral matching at distances greater than 100 meters.",
author = "Hyunju Maeng and Shengcai Liao and Dongoh Kang and Lee, {Seong Whan} and Jain, {Anil K.}",
year = "2013",
month = "4",
day = "11",
doi = "10.1007/978-3-642-37444-9_55",
language = "English",
isbn = "9783642374432",
volume = "7725 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 2",
pages = "708--721",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 2",

}

TY - GEN

T1 - Nighttime face recognition at long distance

T2 - Cross-distance and cross-spectral matching

AU - Maeng, Hyunju

AU - Liao, Shengcai

AU - Kang, Dongoh

AU - Lee, Seong Whan

AU - Jain, Anil K.

PY - 2013/4/11

Y1 - 2013/4/11

N2 - Automatic face recognition capability in surveillance systems is important for security applications. However, few studies have addressed the problem of outdoor face recognition at a long distance (over 100 meters) in both daytime and nighttime environments. In this paper, we first report on a system that we have designed to collect face image database at a long distance, called the Long Distance Heterogeneous Face Database (LDHF-DB) to advance research on this topic. The LDHF-DB contains face images collected in an outdoor environment at distances of 60 meters, 100 meters, and 150 meters, with both visible light (VIS) face images captured in daytime and near infrared (NIR) face images captured in nighttime. Given this database, we have conducted two types of cross-distance face matching (matching long-distance probe to 1-meter gallery) experiments: (i) intra-spectral (VIS to VIS) face matching, and (ii) cross-spectral (NIR to VIS) face matching. The proposed face recognition algorithm consists of following three major steps: (i) Gaussian filtering to remove high frequency noise, (ii) Scale Invariant Feature Transform (SIFT) in local image regions for feature representation, and (iii) a random subspace method to build discriminant subspaces for face recognition. Experimental results show that the proposed face recognition algorithm outperforms two commercial state-of-the-art face recognition SDKs (FaceVACS and PittPatt) for long distance face recognition in both daytime and nighttime operations. These results highlight the need for better data capture setup and robust face matching algorithms for cross spectral matching at distances greater than 100 meters.

AB - Automatic face recognition capability in surveillance systems is important for security applications. However, few studies have addressed the problem of outdoor face recognition at a long distance (over 100 meters) in both daytime and nighttime environments. In this paper, we first report on a system that we have designed to collect face image database at a long distance, called the Long Distance Heterogeneous Face Database (LDHF-DB) to advance research on this topic. The LDHF-DB contains face images collected in an outdoor environment at distances of 60 meters, 100 meters, and 150 meters, with both visible light (VIS) face images captured in daytime and near infrared (NIR) face images captured in nighttime. Given this database, we have conducted two types of cross-distance face matching (matching long-distance probe to 1-meter gallery) experiments: (i) intra-spectral (VIS to VIS) face matching, and (ii) cross-spectral (NIR to VIS) face matching. The proposed face recognition algorithm consists of following three major steps: (i) Gaussian filtering to remove high frequency noise, (ii) Scale Invariant Feature Transform (SIFT) in local image regions for feature representation, and (iii) a random subspace method to build discriminant subspaces for face recognition. Experimental results show that the proposed face recognition algorithm outperforms two commercial state-of-the-art face recognition SDKs (FaceVACS and PittPatt) for long distance face recognition in both daytime and nighttime operations. These results highlight the need for better data capture setup and robust face matching algorithms for cross spectral matching at distances greater than 100 meters.

UR - http://www.scopus.com/inward/record.url?scp=84875911939&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84875911939&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-37444-9_55

DO - 10.1007/978-3-642-37444-9_55

M3 - Conference contribution

AN - SCOPUS:84875911939

SN - 9783642374432

VL - 7725 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 708

EP - 721

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -