Feature fusions for 2.5D face recognition in Random Maxout Extreme Learning Machine

Lee Ying Chong, Thian Song Ong, Beng Jin Teoh

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Contemporary face recognition system is often based on either 2D (texture) or 3D (texture + shape) face modality. An alternative modality that utilizes range (depth) facial images, namely 2.5D face recognition emerges. In this paper, we propose a 2.5D face descriptor that based on the Regional Covariance Matrix (RCM), a powerful means of feature fusion technique and a novel classifier dubbed Random Maxout Extreme Learning Machine (RMELM). The RCM of interest is constructed based on the Principal Component Analysis (PCA) filters responses of facial texture and/or range image, wherein the PCA filters are learned from a two-layer PCA network. The RMELM is an ELM variant where the activation function is based on the locally linear maxout function, in place of typical global non-linear functions in ELM. Since the RCM is a special case of symmetric positive definite matrix that resides on the Tensor manifold; a gap exists in between RCM and RMELM, which is a vector-based classifier. To bridge the gap, we flatten the manifold by transforming the RCM to a feature vector via a matrix logarithm operator. Experimental results from two public 3D face databases, FRGC v2.0 database and Gavab database, validated our proposed method is promising in 2.5D face recognition.

Original languageEnglish
Pages (from-to)358-372
Number of pages15
JournalApplied Soft Computing Journal
Volume75
DOIs
Publication statusPublished - 2019 Feb 1

Fingerprint

Face recognition
Covariance matrix
Learning systems
Fusion reactions
Principal component analysis
Textures
Classifiers
Tensors
Chemical activation

All Science Journal Classification (ASJC) codes

  • Software

Cite this

@article{fd40cb77b7114c6fab694373e89ff7fa,
title = "Feature fusions for 2.5D face recognition in Random Maxout Extreme Learning Machine",
abstract = "Contemporary face recognition system is often based on either 2D (texture) or 3D (texture + shape) face modality. An alternative modality that utilizes range (depth) facial images, namely 2.5D face recognition emerges. In this paper, we propose a 2.5D face descriptor that based on the Regional Covariance Matrix (RCM), a powerful means of feature fusion technique and a novel classifier dubbed Random Maxout Extreme Learning Machine (RMELM). The RCM of interest is constructed based on the Principal Component Analysis (PCA) filters responses of facial texture and/or range image, wherein the PCA filters are learned from a two-layer PCA network. The RMELM is an ELM variant where the activation function is based on the locally linear maxout function, in place of typical global non-linear functions in ELM. Since the RCM is a special case of symmetric positive definite matrix that resides on the Tensor manifold; a gap exists in between RCM and RMELM, which is a vector-based classifier. To bridge the gap, we flatten the manifold by transforming the RCM to a feature vector via a matrix logarithm operator. Experimental results from two public 3D face databases, FRGC v2.0 database and Gavab database, validated our proposed method is promising in 2.5D face recognition.",
author = "Chong, {Lee Ying} and Ong, {Thian Song} and Teoh, {Beng Jin}",
year = "2019",
month = "2",
day = "1",
doi = "10.1016/j.asoc.2018.11.024",
language = "English",
volume = "75",
pages = "358--372",
journal = "Applied Soft Computing",
issn = "1568-4946",
publisher = "Elsevier BV",

}

Feature fusions for 2.5D face recognition in Random Maxout Extreme Learning Machine. / Chong, Lee Ying; Ong, Thian Song; Teoh, Beng Jin.

In: Applied Soft Computing Journal, Vol. 75, 01.02.2019, p. 358-372.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Feature fusions for 2.5D face recognition in Random Maxout Extreme Learning Machine

AU - Chong, Lee Ying

AU - Ong, Thian Song

AU - Teoh, Beng Jin

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Contemporary face recognition system is often based on either 2D (texture) or 3D (texture + shape) face modality. An alternative modality that utilizes range (depth) facial images, namely 2.5D face recognition emerges. In this paper, we propose a 2.5D face descriptor that based on the Regional Covariance Matrix (RCM), a powerful means of feature fusion technique and a novel classifier dubbed Random Maxout Extreme Learning Machine (RMELM). The RCM of interest is constructed based on the Principal Component Analysis (PCA) filters responses of facial texture and/or range image, wherein the PCA filters are learned from a two-layer PCA network. The RMELM is an ELM variant where the activation function is based on the locally linear maxout function, in place of typical global non-linear functions in ELM. Since the RCM is a special case of symmetric positive definite matrix that resides on the Tensor manifold; a gap exists in between RCM and RMELM, which is a vector-based classifier. To bridge the gap, we flatten the manifold by transforming the RCM to a feature vector via a matrix logarithm operator. Experimental results from two public 3D face databases, FRGC v2.0 database and Gavab database, validated our proposed method is promising in 2.5D face recognition.

AB - Contemporary face recognition system is often based on either 2D (texture) or 3D (texture + shape) face modality. An alternative modality that utilizes range (depth) facial images, namely 2.5D face recognition emerges. In this paper, we propose a 2.5D face descriptor that based on the Regional Covariance Matrix (RCM), a powerful means of feature fusion technique and a novel classifier dubbed Random Maxout Extreme Learning Machine (RMELM). The RCM of interest is constructed based on the Principal Component Analysis (PCA) filters responses of facial texture and/or range image, wherein the PCA filters are learned from a two-layer PCA network. The RMELM is an ELM variant where the activation function is based on the locally linear maxout function, in place of typical global non-linear functions in ELM. Since the RCM is a special case of symmetric positive definite matrix that resides on the Tensor manifold; a gap exists in between RCM and RMELM, which is a vector-based classifier. To bridge the gap, we flatten the manifold by transforming the RCM to a feature vector via a matrix logarithm operator. Experimental results from two public 3D face databases, FRGC v2.0 database and Gavab database, validated our proposed method is promising in 2.5D face recognition.

UR - http://www.scopus.com/inward/record.url?scp=85057242678&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057242678&partnerID=8YFLogxK

U2 - 10.1016/j.asoc.2018.11.024

DO - 10.1016/j.asoc.2018.11.024

M3 - Article

AN - SCOPUS:85057242678

VL - 75

SP - 358

EP - 372

JO - Applied Soft Computing

JF - Applied Soft Computing

SN - 1568-4946

ER -