Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; the references it cites are not checked yet.

Lightweight and Resource-Constrained Learning Network for Face Recognition with Performance Optimization

Open access

Abstract

Despite considerable progress in face recognition technology in recent years, deep learning (DL) and convolutional neural networks (CNN) have revealed commendable recognition effects with the advent of artificial intelligence and big data. FaceNet was presented in 2015 and is able to significantly improve the accuracy of face recognition, while also being powerfully built to counteract several common issues, such as occlusion, blur, illumination change, and different angles of head pose. However, not all hardware can sustain the heavy computing load in the execution of the FaceNet model. In applications in the security industry, lightweight and efficient face recognition are two key points for facilitating the deployment of DL and CNN models directly in field devices, due to their limited edge computing capability and low equipment cost. To this end, this paper provides a lightweight learning network improved from FaceNet, which is called FN13, to break through the hardware limitation of constrained computational resources. The proposed FN13 takes the advantage of center loss to reduce the variations of the between-class features and enlarge the difference of the within-class features, instead of the triplet loss by using FaceNet. The resulting model reduces the number of parameters and maintains a high degree of accuracy, only requiring few grayscale reference images per subject. The validity of FN13 is demonstrated by conducting experiments on the Labeled Faces in the Wild (LFW) dataset, as well as an analytical discussion regarding specific disguise problems.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.3390/s20216114

Cites

Statistics

Figures

Figure 1 The architecture of a typical face recognition embedded system on a security device.
Figure 1 The architecture of a typical face recognition embedded system on a security device.CC BY · PubMed Central
Figure 2 Face recognition using FN13 with variation in terms of lighting and postures. The numerical number between the faces represents the similarity of two images where the score ‘zero’ is considered to represent two distinct people.
Figure 2 Face recognition using FN13 with variation in terms of lighting and postures. The numerical number between the faces represents the similarity of two images where the score ‘zero’ is considered to represent two distinct people.CC BY · PubMed Central
Figure 3 Block diagram of FaceNet architecture.
Figure 3 Block diagram of FaceNet architecture.CC BY · PubMed Central
Figure 4 Block diagram of FN13 architecture using center loss function.
Figure 4 Block diagram of FN13 architecture using center loss function.CC BY · PubMed Central
Figure 5 The framework of FN13 model.
Figure 5 The framework of FN13 model.CC BY · PubMed Central
Figure 6 Inception structure modification: (a) Inception-ResNet-A structure with two 3 × 3 convolutions, (b) modified structure based on Inception-ResNet-A, (c) Inception-ResNet-B structure with 1 × 7 and 7 × 1 convolutions, (d) modified structure based on Inception-ResNet-B.
Figure 6 Inception structure modification: (a) Inception-ResNet-A structure with two 3 × 3 convolutions, (b) modified structure based on Inception-ResNet-A, (c) Inception-ResNet-B structure with 1 × 7 and 7 × 1 convolutions, (d) modified structure based on Inception-ResNet-B.CC BY · PubMed Central
Figure 7 Example comparison of features learned under the supervision of softmax loss function with/without center loss introduced to the model. (a) Without center loss, and (b) with center loss; points with different colors represent features for various classes.
Figure 7 Example comparison of features learned under the supervision of softmax loss function with/without center loss introduced to the model. (a) Without center loss, and (b) with center loss; points with different colors represent features for various classes.CC BY · PubMed Central
Figure 8 Image examples selected from the LFW dataset.
Figure 8 Image examples selected from the LFW dataset.CC BY · PubMed Central
Figure 9 The confusion matrix for the face recognition system for evaluating overall performance.
Figure 9 The confusion matrix for the face recognition system for evaluating overall performance.CC BY · PubMed Central
Figure 10 Processing flowchart with FN13 on a real-time face recognition embedded system.
Figure 10 Processing flowchart with FN13 on a real-time face recognition embedded system.CC BY · PubMed Central
Figure 11 Real-time camera recognition of a known face under various conditions (with three samples).
Figure 11 Real-time camera recognition of a known face under various conditions (with three samples).CC BY · PubMed Central
Figure 12 The matrix of similarity examples for the LFW test set with FN13-generated face embedding corresponding to each sample image.
Figure 12 The matrix of similarity examples for the LFW test set with FN13-generated face embedding corresponding to each sample image.CC BY · PubMed Central
Figure 13 Image pair sharing the maximum correlation coefficient in the example of the LFW.
Figure 13 Image pair sharing the maximum correlation coefficient in the example of the LFW.CC BY · PubMed Central
Figure 14 A collection from a self-collected, non-public dataset.
Figure 14 A collection from a self-collected, non-public dataset.CC BY · PubMed Central
Figure 15 Example of falsely accepted face images.
Figure 15 Example of falsely accepted face images.CC BY · PubMed Central
Figure 16 Example of falsely rejected face images.
Figure 16 Example of falsely rejected face images.CC BY · PubMed Central
Figure 17 The matrix of similarity for the falsely accepted face images shown in Figure 16.
Figure 17 The matrix of similarity for the falsely accepted face images shown in Figure 16.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

The citation lists are queued

They are fetched later and appear here then.

References 37 · Cited by 54

Metadata

Authors
Li, H.-C., Deng, Z.-Y., & Chiang, H.-H.
Affiliations
  • Fu Jen Catholic University (TW)
  • National Taiwan Normal University (TW)
Corresponding author
Chiang, Hsin-Han
Venue
Sensors
Publisher
MDPI AG
Published
2020-10-27
Volume
20
Issue
21
Pages
6114
ISSN
1424-8220 (ISSN-L)
OpenAlex
W3097781980
Cited by
54
References
37
Citations by year
2026: 5 · 2025: 11 · 2024: 17 · 2023: 14 · 2022: 5 · 2021: 2
FWCI
2.04
Type
journal article
Language
en
Open access
Locations
  • Sensors (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Holtek Semiconductor (Taiwan) Inc. (108K0007)
  • Ministry of Science and Technology, Taiwan, R.O.C. (MOST 107-2221-E-003-024-MY3 and MOST 109-2634-F-003-006)
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face and Expression Recognition
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
Keywords
face recognition, lightweight neural networks, FaceNet, center loss, triplet loss, edge computing, resource-constrained devices, parameter reduction, LFW dataset
MeSH terms
Deep Learning, Automated Facial Recognition, Algorithms, Artificial Intelligence, Head, Humans, Neural Networks, Computer

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Lightweight and Resource-Constrained Learning Network for Face Recognition with Performance Optimization. https://app.cochat.ai/papers/10.3390/s20216114
The paper
Li, H.-C., Deng, Z.-Y., & Chiang, H.-H. (2020). Lightweight and Resource-Constrained Learning Network for Face Recognition with Performance Optimization. Sensors. https://doi.org/10.3390/s20216114

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.