FaceNet with RetinaFace to Identify Masked Face

Sengli Egani Sitepu; Grafika Jati; Machmud Roby Alhamidi; Wahyu Caesarendra; Wisnu Jatmiko

doi:10.1109/IWBIS53353.2021.9631848

FaceNet with RetinaFace to Identify Masked Face

Sengli Egani Sitepu, Grafika Jati, Machmud Roby Alhamidi, Wahyu Caesarendra, Wisnu Jatmiko

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

5 Citations (Scopus)

Abstract

The use of masks due to the Covid-19 pandemic reduces the accuracy of facial recognition systems applied to camera-based security systems. The use of the mask by the people covers most of the facial featureswhich is located from middle to bottom area. In addition, the area which are still visible are the upper face which are eyes and forehead. This paper proposes a masked face recognition using a combination of RetinaFace as a face detector and FaceNet as a face recognizer. The MFR2 dataset with 53 identities was used to train and test this method. The test data in this study are only images of masked faces. Cosine Distance was implemented to measure the face similarity. Based on the experiment results, the proposed method obtained 98.2% of detection accuracy. The proposed method provided 78% accurate performance with 3.63 s for processing a single frame in terms of face recognition. The performance indicates that our system can potentially be applied in security systems with many different identities.

Original language	English
Title of host publication	Proceedings - IWBIS 2021
Subtitle of host publication	6th International Workshop on Big Data and Information Security
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	81-86
Number of pages	6
ISBN (Electronic)	9781665424516
DOIs	https://doi.org/10.1109/IWBIS53353.2021.9631848
Publication status	Published - 2021
Event	6th International Workshop on Big Data and Information Security, IWBIS 2021 - Virtual, Online, Indonesia Duration: 23 Oct 2021 → 26 Oct 2021

Publication series

Name	Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security

Conference

Conference	6th International Workshop on Big Data and Information Security, IWBIS 2021
Country/Territory	Indonesia
City	Virtual, Online
Period	23/10/21 → 26/10/21

Keywords

face detection
face recognition
FaceNet
masked face
RetinaFace

Access to Document

10.1109/IWBIS53353.2021.9631848

Cite this

Sitepu, S. E., Jati, G., Alhamidi, M. R., Caesarendra, W., & Jatmiko, W. (2021). FaceNet with RetinaFace to Identify Masked Face. In Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security (pp. 81-86). (Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IWBIS53353.2021.9631848

Sitepu, Sengli Egani ; Jati, Grafika ; Alhamidi, Machmud Roby et al. / FaceNet with RetinaFace to Identify Masked Face. Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security. Institute of Electrical and Electronics Engineers Inc., 2021. pp. 81-86 (Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security).

@inproceedings{fbefd2d55235456ea6f1bbea831a8d75,

title = "FaceNet with RetinaFace to Identify Masked Face",

abstract = "The use of masks due to the Covid-19 pandemic reduces the accuracy of facial recognition systems applied to camera-based security systems. The use of the mask by the people covers most of the facial featureswhich is located from middle to bottom area. In addition, the area which are still visible are the upper face which are eyes and forehead. This paper proposes a masked face recognition using a combination of RetinaFace as a face detector and FaceNet as a face recognizer. The MFR2 dataset with 53 identities was used to train and test this method. The test data in this study are only images of masked faces. Cosine Distance was implemented to measure the face similarity. Based on the experiment results, the proposed method obtained 98.2% of detection accuracy. The proposed method provided 78% accurate performance with 3.63 s for processing a single frame in terms of face recognition. The performance indicates that our system can potentially be applied in security systems with many different identities. ",

keywords = "face detection, face recognition, FaceNet, masked face, RetinaFace",

author = "Sitepu, {Sengli Egani} and Grafika Jati and Alhamidi, {Machmud Roby} and Wahyu Caesarendra and Wisnu Jatmiko",

note = "Funding Information: ACKNOWLEDGEMENT This works was supported by Konsorsium Riset Unggulan Perguruan Tinggi (KRUPT) from National Research and Innovation Agency entitled “Pengembangan Integrated Smart Portable Telehealth USG System untuk Meningkatkan Indeks Kesehatan Ibu dan Bayi di Indonesia”, NKB-390/ UN2.RST/HKP.05.00/2021. Publisher Copyright: {\textcopyright} 2021 IEEE.; 6th International Workshop on Big Data and Information Security, IWBIS 2021 ; Conference date: 23-10-2021 Through 26-10-2021",

year = "2021",

doi = "10.1109/IWBIS53353.2021.9631848",

language = "English",

series = "Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "81--86",

booktitle = "Proceedings - IWBIS 2021",

address = "United States",

}

Sitepu, SE, Jati, G, Alhamidi, MR, Caesarendra, W & Jatmiko, W 2021, FaceNet with RetinaFace to Identify Masked Face. in Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security. Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security, Institute of Electrical and Electronics Engineers Inc., pp. 81-86, 6th International Workshop on Big Data and Information Security, IWBIS 2021, Virtual, Online, Indonesia, 23/10/21. https://doi.org/10.1109/IWBIS53353.2021.9631848

FaceNet with RetinaFace to Identify Masked Face. / Sitepu, Sengli Egani; Jati, Grafika; Alhamidi, Machmud Roby et al.
Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security. Institute of Electrical and Electronics Engineers Inc., 2021. p. 81-86 (Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - FaceNet with RetinaFace to Identify Masked Face

AU - Sitepu, Sengli Egani

AU - Jati, Grafika

AU - Alhamidi, Machmud Roby

AU - Caesarendra, Wahyu

AU - Jatmiko, Wisnu

N1 - Funding Information: ACKNOWLEDGEMENT This works was supported by Konsorsium Riset Unggulan Perguruan Tinggi (KRUPT) from National Research and Innovation Agency entitled “Pengembangan Integrated Smart Portable Telehealth USG System untuk Meningkatkan Indeks Kesehatan Ibu dan Bayi di Indonesia”, NKB-390/ UN2.RST/HKP.05.00/2021. Publisher Copyright: © 2021 IEEE.

PY - 2021

Y1 - 2021

N2 - The use of masks due to the Covid-19 pandemic reduces the accuracy of facial recognition systems applied to camera-based security systems. The use of the mask by the people covers most of the facial featureswhich is located from middle to bottom area. In addition, the area which are still visible are the upper face which are eyes and forehead. This paper proposes a masked face recognition using a combination of RetinaFace as a face detector and FaceNet as a face recognizer. The MFR2 dataset with 53 identities was used to train and test this method. The test data in this study are only images of masked faces. Cosine Distance was implemented to measure the face similarity. Based on the experiment results, the proposed method obtained 98.2% of detection accuracy. The proposed method provided 78% accurate performance with 3.63 s for processing a single frame in terms of face recognition. The performance indicates that our system can potentially be applied in security systems with many different identities.

AB - The use of masks due to the Covid-19 pandemic reduces the accuracy of facial recognition systems applied to camera-based security systems. The use of the mask by the people covers most of the facial featureswhich is located from middle to bottom area. In addition, the area which are still visible are the upper face which are eyes and forehead. This paper proposes a masked face recognition using a combination of RetinaFace as a face detector and FaceNet as a face recognizer. The MFR2 dataset with 53 identities was used to train and test this method. The test data in this study are only images of masked faces. Cosine Distance was implemented to measure the face similarity. Based on the experiment results, the proposed method obtained 98.2% of detection accuracy. The proposed method provided 78% accurate performance with 3.63 s for processing a single frame in terms of face recognition. The performance indicates that our system can potentially be applied in security systems with many different identities.

KW - face detection

KW - face recognition

KW - FaceNet

KW - masked face

KW - RetinaFace

UR - http://www.scopus.com/inward/record.url?scp=85124360801&partnerID=8YFLogxK

U2 - 10.1109/IWBIS53353.2021.9631848

DO - 10.1109/IWBIS53353.2021.9631848

M3 - Conference contribution

AN - SCOPUS:85124360801

T3 - Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security

SP - 81

EP - 86

BT - Proceedings - IWBIS 2021

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 6th International Workshop on Big Data and Information Security, IWBIS 2021

Y2 - 23 October 2021 through 26 October 2021

ER -

Sitepu SE, Jati G, Alhamidi MR, Caesarendra W, Jatmiko W. FaceNet with RetinaFace to Identify Masked Face. In Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security. Institute of Electrical and Electronics Engineers Inc. 2021. p. 81-86. (Proceedings - IWBIS 2021: 6th International Workshop on Big Data and Information Security). doi: 10.1109/IWBIS53353.2021.9631848

FaceNet with RetinaFace to Identify Masked Face

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this