Face recognition using spectral and spatial information

Stefan Robila, Marco Chang, Nisha B. D'Amico

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

We present a novel unsupervised method for facial recognition using hyperspectral imaging and decision fusion. In previous work we have separately investigated the use of spectra matching and image based matching. In spectra matching, face spectra are being classified based on spectral similarities. In image based matching, we investigated various approaches based on orthogonal subspaces (such as PCA and OSP). In the current work we provide an automated unsupervised method that starts by detecting the face in the image and then proceeds to performs both spectral and image based matching. The results are fused in a single classification decision. The algorithm is tested on an experimental hyperspectral image database of 17 subjects each with five different facial expressions and viewing angles. Our results show that the decision fusion leads to improvement of recognition accuracy when compared to the individual approaches as well as to recognition based on regular imaging.

Original languageEnglish
Title of host publicationApplications of Digital Image Processing XXXIV
Volume8185
DOIs
StatePublished - 31 Oct 2011
EventApplications of Digital Image Processing XXXIV - San Diego, CA, United States
Duration: 22 Aug 201124 Aug 2011

Other

OtherApplications of Digital Image Processing XXXIV
CountryUnited States
CitySan Diego, CA
Period22/08/1124/08/11

Fingerprint

Spatial Information
Face recognition
Face Recognition
Fusion reactions
Decision Fusion
Face
Imaging techniques
Hyperspectral Imaging
fusion
Hyperspectral Image
Facial Expression
Image Database
Subspace
Imaging
Angle
Hyperspectral imaging

Keywords

  • face recognition
  • hyperspectral imagery
  • orthogonal subspace projection

Cite this

Robila, S., Chang, M., & D'Amico, N. B. (2011). Face recognition using spectral and spatial information. In Applications of Digital Image Processing XXXIV (Vol. 8185). [81351Q] https://doi.org/10.1117/12.892743
Robila, Stefan ; Chang, Marco ; D'Amico, Nisha B. / Face recognition using spectral and spatial information. Applications of Digital Image Processing XXXIV. Vol. 8185 2011.
@inproceedings{28acaa7d905a4392943720379605a78a,
title = "Face recognition using spectral and spatial information",
abstract = "We present a novel unsupervised method for facial recognition using hyperspectral imaging and decision fusion. In previous work we have separately investigated the use of spectra matching and image based matching. In spectra matching, face spectra are being classified based on spectral similarities. In image based matching, we investigated various approaches based on orthogonal subspaces (such as PCA and OSP). In the current work we provide an automated unsupervised method that starts by detecting the face in the image and then proceeds to performs both spectral and image based matching. The results are fused in a single classification decision. The algorithm is tested on an experimental hyperspectral image database of 17 subjects each with five different facial expressions and viewing angles. Our results show that the decision fusion leads to improvement of recognition accuracy when compared to the individual approaches as well as to recognition based on regular imaging.",
keywords = "face recognition, hyperspectral imagery, orthogonal subspace projection",
author = "Stefan Robila and Marco Chang and D'Amico, {Nisha B.}",
year = "2011",
month = "10",
day = "31",
doi = "10.1117/12.892743",
language = "English",
isbn = "9780819487452",
volume = "8185",
booktitle = "Applications of Digital Image Processing XXXIV",

}

Robila, S, Chang, M & D'Amico, NB 2011, Face recognition using spectral and spatial information. in Applications of Digital Image Processing XXXIV. vol. 8185, 81351Q, Applications of Digital Image Processing XXXIV, San Diego, CA, United States, 22/08/11. https://doi.org/10.1117/12.892743

Face recognition using spectral and spatial information. / Robila, Stefan; Chang, Marco; D'Amico, Nisha B.

Applications of Digital Image Processing XXXIV. Vol. 8185 2011. 81351Q.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

TY - GEN

T1 - Face recognition using spectral and spatial information

AU - Robila, Stefan

AU - Chang, Marco

AU - D'Amico, Nisha B.

PY - 2011/10/31

Y1 - 2011/10/31

N2 - We present a novel unsupervised method for facial recognition using hyperspectral imaging and decision fusion. In previous work we have separately investigated the use of spectra matching and image based matching. In spectra matching, face spectra are being classified based on spectral similarities. In image based matching, we investigated various approaches based on orthogonal subspaces (such as PCA and OSP). In the current work we provide an automated unsupervised method that starts by detecting the face in the image and then proceeds to performs both spectral and image based matching. The results are fused in a single classification decision. The algorithm is tested on an experimental hyperspectral image database of 17 subjects each with five different facial expressions and viewing angles. Our results show that the decision fusion leads to improvement of recognition accuracy when compared to the individual approaches as well as to recognition based on regular imaging.

AB - We present a novel unsupervised method for facial recognition using hyperspectral imaging and decision fusion. In previous work we have separately investigated the use of spectra matching and image based matching. In spectra matching, face spectra are being classified based on spectral similarities. In image based matching, we investigated various approaches based on orthogonal subspaces (such as PCA and OSP). In the current work we provide an automated unsupervised method that starts by detecting the face in the image and then proceeds to performs both spectral and image based matching. The results are fused in a single classification decision. The algorithm is tested on an experimental hyperspectral image database of 17 subjects each with five different facial expressions and viewing angles. Our results show that the decision fusion leads to improvement of recognition accuracy when compared to the individual approaches as well as to recognition based on regular imaging.

KW - face recognition

KW - hyperspectral imagery

KW - orthogonal subspace projection

UR - http://www.scopus.com/inward/record.url?scp=80054915912&partnerID=8YFLogxK

U2 - 10.1117/12.892743

DO - 10.1117/12.892743

M3 - Conference contribution

SN - 9780819487452

VL - 8185

BT - Applications of Digital Image Processing XXXIV

ER -

Robila S, Chang M, D'Amico NB. Face recognition using spectral and spatial information. In Applications of Digital Image Processing XXXIV. Vol. 8185. 2011. 81351Q https://doi.org/10.1117/12.892743