Adapting local features for face detection in thermal image

Chao Ma, Ngo Thanh Trung, Hideaki Uchiyama, Hajime Nagahara, Atsushi Shimada, Rin Ichiro Taniguchi

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

A thermal camera captures the temperature distribution of a scene as a thermal image. In thermal images, facial appearances of different people under different lighting conditions are similar. This is because facial temperature distribution is generally constant and not affected by lighting condition. This similarity in face appearances is advantageous for face detection. To detect faces in thermal images, cascade classifiers with Haar-like features are generally used. However, there are few studies exploring the local features for face detection in thermal images. In this paper, we introduce two approaches relying on local features for face detection in thermal images. First, we create new feature types by extending Multi-Block LBP. We consider a margin around the reference and the generally constant distribution of facial temperature. In this way, we make the features more robust to image noise and more effective for face detection in thermal images. Second, we propose an AdaBoost-based training method to get cascade classifiers with multiple types of local features. These feature types have different advantages. In this way we enhance the description power of local features. We did a hold-out validation experiment and a field experiment. In the hold-out validation experiment, we captured a dataset from 20 participants, comprising 14 males and 6 females. For each participant, we captured 420 images with 10 variations in camera distance, 21 poses, and 2 appearances (participant with/without glasses). We compared the performance of cascade classifiers trained by different sets of the features. The experiment results showed that the proposed approaches effectively improve the performance of face detection in thermal images. In the field experiment, we compared the face detection performance in realistic scenes using thermal and RGB images, and gave discussion based on the results.

Original languageEnglish
Article number2741
JournalSensors (Switzerland)
Volume17
Issue number12
DOIs
Publication statusPublished - Dec 2017

Fingerprint

Face recognition
Hot Temperature
classifiers
Classifiers
cascades
Lighting
Experiments
Temperature
illuminating
Temperature distribution
Cameras
temperature distribution
cameras
Adaptive boosting
Glass
margins
education
glass

All Science Journal Classification (ASJC) codes

  • Analytical Chemistry
  • Atomic and Molecular Physics, and Optics
  • Biochemistry
  • Instrumentation
  • Electrical and Electronic Engineering

Cite this

Adapting local features for face detection in thermal image. / Ma, Chao; Trung, Ngo Thanh; Uchiyama, Hideaki; Nagahara, Hajime; Shimada, Atsushi; Taniguchi, Rin Ichiro.

In: Sensors (Switzerland), Vol. 17, No. 12, 2741, 12.2017.

Research output: Contribution to journalArticle

@article{4292dd51959c4d38b8110253494ad18d,
title = "Adapting local features for face detection in thermal image",
abstract = "A thermal camera captures the temperature distribution of a scene as a thermal image. In thermal images, facial appearances of different people under different lighting conditions are similar. This is because facial temperature distribution is generally constant and not affected by lighting condition. This similarity in face appearances is advantageous for face detection. To detect faces in thermal images, cascade classifiers with Haar-like features are generally used. However, there are few studies exploring the local features for face detection in thermal images. In this paper, we introduce two approaches relying on local features for face detection in thermal images. First, we create new feature types by extending Multi-Block LBP. We consider a margin around the reference and the generally constant distribution of facial temperature. In this way, we make the features more robust to image noise and more effective for face detection in thermal images. Second, we propose an AdaBoost-based training method to get cascade classifiers with multiple types of local features. These feature types have different advantages. In this way we enhance the description power of local features. We did a hold-out validation experiment and a field experiment. In the hold-out validation experiment, we captured a dataset from 20 participants, comprising 14 males and 6 females. For each participant, we captured 420 images with 10 variations in camera distance, 21 poses, and 2 appearances (participant with/without glasses). We compared the performance of cascade classifiers trained by different sets of the features. The experiment results showed that the proposed approaches effectively improve the performance of face detection in thermal images. In the field experiment, we compared the face detection performance in realistic scenes using thermal and RGB images, and gave discussion based on the results.",
author = "Chao Ma and Trung, {Ngo Thanh} and Hideaki Uchiyama and Hajime Nagahara and Atsushi Shimada and Taniguchi, {Rin Ichiro}",
year = "2017",
month = "12",
doi = "10.3390/s17122741",
language = "English",
volume = "17",
journal = "Sensors",
issn = "1424-3210",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "12",

}

TY - JOUR

T1 - Adapting local features for face detection in thermal image

AU - Ma, Chao

AU - Trung, Ngo Thanh

AU - Uchiyama, Hideaki

AU - Nagahara, Hajime

AU - Shimada, Atsushi

AU - Taniguchi, Rin Ichiro

PY - 2017/12

Y1 - 2017/12

N2 - A thermal camera captures the temperature distribution of a scene as a thermal image. In thermal images, facial appearances of different people under different lighting conditions are similar. This is because facial temperature distribution is generally constant and not affected by lighting condition. This similarity in face appearances is advantageous for face detection. To detect faces in thermal images, cascade classifiers with Haar-like features are generally used. However, there are few studies exploring the local features for face detection in thermal images. In this paper, we introduce two approaches relying on local features for face detection in thermal images. First, we create new feature types by extending Multi-Block LBP. We consider a margin around the reference and the generally constant distribution of facial temperature. In this way, we make the features more robust to image noise and more effective for face detection in thermal images. Second, we propose an AdaBoost-based training method to get cascade classifiers with multiple types of local features. These feature types have different advantages. In this way we enhance the description power of local features. We did a hold-out validation experiment and a field experiment. In the hold-out validation experiment, we captured a dataset from 20 participants, comprising 14 males and 6 females. For each participant, we captured 420 images with 10 variations in camera distance, 21 poses, and 2 appearances (participant with/without glasses). We compared the performance of cascade classifiers trained by different sets of the features. The experiment results showed that the proposed approaches effectively improve the performance of face detection in thermal images. In the field experiment, we compared the face detection performance in realistic scenes using thermal and RGB images, and gave discussion based on the results.

AB - A thermal camera captures the temperature distribution of a scene as a thermal image. In thermal images, facial appearances of different people under different lighting conditions are similar. This is because facial temperature distribution is generally constant and not affected by lighting condition. This similarity in face appearances is advantageous for face detection. To detect faces in thermal images, cascade classifiers with Haar-like features are generally used. However, there are few studies exploring the local features for face detection in thermal images. In this paper, we introduce two approaches relying on local features for face detection in thermal images. First, we create new feature types by extending Multi-Block LBP. We consider a margin around the reference and the generally constant distribution of facial temperature. In this way, we make the features more robust to image noise and more effective for face detection in thermal images. Second, we propose an AdaBoost-based training method to get cascade classifiers with multiple types of local features. These feature types have different advantages. In this way we enhance the description power of local features. We did a hold-out validation experiment and a field experiment. In the hold-out validation experiment, we captured a dataset from 20 participants, comprising 14 males and 6 females. For each participant, we captured 420 images with 10 variations in camera distance, 21 poses, and 2 appearances (participant with/without glasses). We compared the performance of cascade classifiers trained by different sets of the features. The experiment results showed that the proposed approaches effectively improve the performance of face detection in thermal images. In the field experiment, we compared the face detection performance in realistic scenes using thermal and RGB images, and gave discussion based on the results.

UR - http://www.scopus.com/inward/record.url?scp=85036458665&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85036458665&partnerID=8YFLogxK

U2 - 10.3390/s17122741

DO - 10.3390/s17122741

M3 - Article

C2 - 29186923

AN - SCOPUS:85036458665

VL - 17

JO - Sensors

JF - Sensors

SN - 1424-3210

IS - 12

M1 - 2741

ER -