Finessing filter scarcity problem in face recognition via multi-fold filter convolution

Cheng Yaw Low, Beng Jin Teoh

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

1 Citation (Scopus)

Abstract

The deep convolutional neural networks for face recognition, from DeepFace to the recent FaceNet, demand a sufficiently large volume of filters for feature extraction, in addition to being deep. The shallow filter-bank approaches, e.g., principal component analysis network (PCANet), binarized statistical image features (BSIF), and other analogous variants, endure the filter scarcity problem that not all PCA and ICA filters available are discriminative to abstract noise-free features. This paper extends our previous work on multi-fold filter convolution (â.,3-FFC), where the pre-learned PCA and ICA filter sets are exponentially diversified by â.,3 folds to instantiate PCA, ICA, and PCA-ICA offspring. The experimental results unveil that the 2-FFC operation solves the filter scarcity state. The 2-FFC descriptors are also evidenced to be superior to that of PCANet, BSIF, and other face descriptors, in terms of rank-1 identification rate (%).

Original languageEnglish
Title of host publicationSecond International Workshop on Pattern Recognition
EditorsGuojian Chen, Xudong Jiang, Masayuki Arai
PublisherSPIE
Volume10443
ISBN (Electronic)9781510613508
DOIs
Publication statusPublished - 2017 Jan 1
Event2nd International Workshop on Pattern Recognition, IWPR 2017 - Singapore, Singapore
Duration: 2017 May 12017 May 3

Other

Other2nd International Workshop on Pattern Recognition, IWPR 2017
CountrySingapore
CitySingapore
Period17/5/117/5/3

Fingerprint

Independent component analysis
Face recognition
Face Recognition
Convolution
convolution integrals
Fold
Filter
filters
Principal component analysis
Principal Component Analysis
Descriptors
principal components analysis
Filter banks
Feature extraction
Filter Banks
Neural networks
Feature Extraction
finesse
pattern recognition
Face

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Cite this

Low, C. Y., & Teoh, B. J. (2017). Finessing filter scarcity problem in face recognition via multi-fold filter convolution. In G. Chen, X. Jiang, & M. Arai (Eds.), Second International Workshop on Pattern Recognition (Vol. 10443). [104430G] SPIE. https://doi.org/10.1117/12.2280352
Low, Cheng Yaw ; Teoh, Beng Jin. / Finessing filter scarcity problem in face recognition via multi-fold filter convolution. Second International Workshop on Pattern Recognition. editor / Guojian Chen ; Xudong Jiang ; Masayuki Arai. Vol. 10443 SPIE, 2017.
@inproceedings{a9804b19aeec403181e000fda0790a7c,
title = "Finessing filter scarcity problem in face recognition via multi-fold filter convolution",
abstract = "The deep convolutional neural networks for face recognition, from DeepFace to the recent FaceNet, demand a sufficiently large volume of filters for feature extraction, in addition to being deep. The shallow filter-bank approaches, e.g., principal component analysis network (PCANet), binarized statistical image features (BSIF), and other analogous variants, endure the filter scarcity problem that not all PCA and ICA filters available are discriminative to abstract noise-free features. This paper extends our previous work on multi-fold filter convolution ({\^a}.,3-FFC), where the pre-learned PCA and ICA filter sets are exponentially diversified by {\^a}.,3 folds to instantiate PCA, ICA, and PCA-ICA offspring. The experimental results unveil that the 2-FFC operation solves the filter scarcity state. The 2-FFC descriptors are also evidenced to be superior to that of PCANet, BSIF, and other face descriptors, in terms of rank-1 identification rate ({\%}).",
author = "Low, {Cheng Yaw} and Teoh, {Beng Jin}",
year = "2017",
month = "1",
day = "1",
doi = "10.1117/12.2280352",
language = "English",
volume = "10443",
editor = "Guojian Chen and Xudong Jiang and Masayuki Arai",
booktitle = "Second International Workshop on Pattern Recognition",
publisher = "SPIE",
address = "United States",

}

Low, CY & Teoh, BJ 2017, Finessing filter scarcity problem in face recognition via multi-fold filter convolution. in G Chen, X Jiang & M Arai (eds), Second International Workshop on Pattern Recognition. vol. 10443, 104430G, SPIE, 2nd International Workshop on Pattern Recognition, IWPR 2017, Singapore, Singapore, 17/5/1. https://doi.org/10.1117/12.2280352

Finessing filter scarcity problem in face recognition via multi-fold filter convolution. / Low, Cheng Yaw; Teoh, Beng Jin.

Second International Workshop on Pattern Recognition. ed. / Guojian Chen; Xudong Jiang; Masayuki Arai. Vol. 10443 SPIE, 2017. 104430G.

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

TY - GEN

T1 - Finessing filter scarcity problem in face recognition via multi-fold filter convolution

AU - Low, Cheng Yaw

AU - Teoh, Beng Jin

PY - 2017/1/1

Y1 - 2017/1/1

N2 - The deep convolutional neural networks for face recognition, from DeepFace to the recent FaceNet, demand a sufficiently large volume of filters for feature extraction, in addition to being deep. The shallow filter-bank approaches, e.g., principal component analysis network (PCANet), binarized statistical image features (BSIF), and other analogous variants, endure the filter scarcity problem that not all PCA and ICA filters available are discriminative to abstract noise-free features. This paper extends our previous work on multi-fold filter convolution (â.,3-FFC), where the pre-learned PCA and ICA filter sets are exponentially diversified by â.,3 folds to instantiate PCA, ICA, and PCA-ICA offspring. The experimental results unveil that the 2-FFC operation solves the filter scarcity state. The 2-FFC descriptors are also evidenced to be superior to that of PCANet, BSIF, and other face descriptors, in terms of rank-1 identification rate (%).

AB - The deep convolutional neural networks for face recognition, from DeepFace to the recent FaceNet, demand a sufficiently large volume of filters for feature extraction, in addition to being deep. The shallow filter-bank approaches, e.g., principal component analysis network (PCANet), binarized statistical image features (BSIF), and other analogous variants, endure the filter scarcity problem that not all PCA and ICA filters available are discriminative to abstract noise-free features. This paper extends our previous work on multi-fold filter convolution (â.,3-FFC), where the pre-learned PCA and ICA filter sets are exponentially diversified by â.,3 folds to instantiate PCA, ICA, and PCA-ICA offspring. The experimental results unveil that the 2-FFC operation solves the filter scarcity state. The 2-FFC descriptors are also evidenced to be superior to that of PCANet, BSIF, and other face descriptors, in terms of rank-1 identification rate (%).

UR - http://www.scopus.com/inward/record.url?scp=85028548855&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028548855&partnerID=8YFLogxK

U2 - 10.1117/12.2280352

DO - 10.1117/12.2280352

M3 - Conference contribution

AN - SCOPUS:85028548855

VL - 10443

BT - Second International Workshop on Pattern Recognition

A2 - Chen, Guojian

A2 - Jiang, Xudong

A2 - Arai, Masayuki

PB - SPIE

ER -

Low CY, Teoh BJ. Finessing filter scarcity problem in face recognition via multi-fold filter convolution. In Chen G, Jiang X, Arai M, editors, Second International Workshop on Pattern Recognition. Vol. 10443. SPIE. 2017. 104430G https://doi.org/10.1117/12.2280352