See discussions, stats, and author profiles for this publication at: [Link]

net/publication/361476376

Face Recognition Using MATLAB

Conference Paper · December 2021

DOI: 10.1109/CSCI54926.2021.00331

CITATION READS

1 244

3 authors, including:

Ibidun Christiana Obagbuwa

Sol Plaatje University
81 PUBLICATIONS 814 CITATIONS

SEE PROFILE

All content following this page was uploaded by Ibidun Christiana Obagbuwa on 20 October 2024.

The user has requested enhancement of the downloaded file.

 FACE RECOGNITION Using MATLAB
Ibidun Christiana Obagbuwa Stefany Bam
Department of Computer Science and Information Technology Department of Computer Science and Information Technology
Sol Plaatie University Sol Plaatie University
Kimberley, South Africa Kimberley, South Africa
[Link]@[Link] 201900419@[Link]

Dineo Tiffany Moroka

Department of Computer Science and Information Technology
Sol Plaatie University
Kimberley, South Africa
201902565@[Link]

Abstract - This work sought to investigate how face recognition

can be implemented in MATLAB to correctly detect and identify
an individual using their face. Face recognition is a biometric
technology that is used to recognize and authenticate a II. STATEMENT OF THE PROBLEM
detected face in images or videos. This technology can be
used in various industries for various purposes. The main
goal of this work is to correctly authenticate an individual The protection of information and data is very important,
face using Convolutional Neural Networks (called AlexNet) and a secure security system plays a vital role in it.
in MATLAB.
Face recognition can be used to accurately identify,
Keywords – Face recognition, Biometric Technology, MATLAB, interdict, and apprehend criminal suspects. It also aids in
AlexNet, CNN.
protecting personal data from crimes such as fraud.

I. INTRODUCTION
III. SIGNIFICANCE OF THE PROBLEM
Face recognition is a frictionless biometric that allows
The main aspect of this facial recognition system is to
for the recognition and identification of human faces
correctly identify and verify an individual. The system
based on biometric indicators found within the face that
can be applied in many industries such as law
are unique to an individual [1].
enforcement, banking, border control, and healthcare for
safety and security.
It works by capturing, analyzing, and comparing patterns
based on facial features (spacing of the eyes, bulging of
the nose, the contour of the lips, ears, chin, etc.). The face IV. AIM
recognition algorithm is mainly focused on, face
detection (detecting and locating human faces), face
capturing (converting analog information into digital This project aims to create a system using MATLAB that
information), and face match (verifying if faces belong can accurately verify and identify any individual.
to the same person). V. OBJECTIVES
The era of facial recognition started back in the 1960s. Woody
Bledsoe, Helen Chan Wolf, and Charles Bisson worked on 1. To create our dataset using MATLAB that will
manually marking various “landmarks” on an be used to implement Face recognition.
individual’s face. These “landmarks” included the mouth,
nose, eyes, and ears. The distances between “landmarks” were 2. To detect a face and return a 227×227 image of
calculated and then used to compare images to determine the face.
identity
[Link] 3. To train a Convolutional Neural Networks
(CNN)-AlexNet in MATLAB using the labeled
dataset we collected of the faces and obtain a
Today, face recognition plays a very significant role in Series network which is a pre-trained network
safety and security. It is one of the most preferred [2-7].
biometric technologies because it is easy to set up and
implement and it is user-friendly.
 4. To use the Series network in a testing model to C. Methods of Data analysis
authenticate the test individually.
1. Input image/ webcam video of the face.
VI. LIMITATIONS OF THE STUDY 2. Face and/or facial landmarks are detected in the
image or each video frame.
Errors such as a false negative or false positive
may arise due to poor lighting, bad angles, or 3. Pre-processing is performed on the image or
low-quality image resolution. video frames which includes detecting the face
in the image/ video frame using
[Link] and then to return
VII. LITERATURE REVIEW the detected face as a 227×227 image. The image
is then saved to a remote labeled file on the
1. Techniques and Challenges of Face laptop. The process is repeated for each image/
Recognition: A Critical Review, by Shilpi video 150 times [11].
Singh, S.V.A.V. Prasad focused on reviewing
face recognition techniques and their challenges 4. The gathered images in their respected labeled
for accuracy, recognition rate, and performance folders are placed into a database for model
comparison on different face databases. Aging, training.
pose variation and expression were some of the
challenges considered [8]. 5. We feed our database through a CNN called
AlexNet with a max number of epochs at 20.
2. Face recognition: Past, present, and future (a
review), by Murat Taskiran, Nihan Kahraman, 6. The resulting pre-trained layers are then saved
and Cigdem Eroglu, focused on reviewing and into a series network.
comparing image and video-based face
recognition methods. The authors also discussed 7. The series network is used in the test model and
the evolution of face recognition as well as the imports test images to perform face detection
main steps of a traditional face recognition and authentication.
system [9]

3. AlexNet: The Architecture that Challenged CNNs by

8. Lastly, a label image using a series network is
Jerry Wei, discussed the need for capable deep
returned.
models such as AlexNet to handle large datasets.
Features such as ReLu Nonlinearity, Multiple GUPs,
and Overlapping Pooling were also discussed and
said to be the new approach to convolutional neural XI. FACE RECOGNITION RESULTS
networks [2].

4. Biometrics A Look at Facial Recognition by John D.

Woodward, Jr., Christopher Horn, Julius Gatune, and
Aryn Thomas explores the literature of biometrics
and focuses on face recognition in-depth as well as
the legal status quo related to it [8, 10].

VIII. METHODOLOGY

A. Method of collection

The data used for this project was obtained by asking for
permission from fellow students to use their various
images. All the people in our dataset gave their consent
and we only labeled our data with their names to give
them some form of privacy.

B. System specifications Figure 1: Face recognition Output

MATLAB R2021a was used to implement this system.

In MATLAB R2021a we also used the MATLAB
Support Package for USB Webcams, Computer Vision
Toolbox Interface for Open CV in MATLAB, and Deep
Learning Toolbox Model for AlexNet Network Add-Ons
[3-7].
 XII. PROBLEMS EXPERIENCED
The system created yielded some errors due to poor
lighting, odd angles, and shadings.

Figure 2: Face recognition Output Figure 5: Face recognition Output that detected
no face

Figure 6 shows an output of image

misclassification, the system was not able to
classify the image correctly.

Figure 3: Face recognition Output Figure 6: Image misclassification Face

recognition Output

XIII. CONCLUSION
In this project, face recognition was implemented in
MATLAB by using a convolution network named
AlexNet. AlexNet has 8 deep layers where five of which
are convolutional layers and three of them are fully
connected layers.

The system created can collect face images using the

webcam and this can be helpful for real-time data
collection as it plays a role in safety and security.

Although errors such as misclassification and faces not

being detected appeared due to poor lighting, odd angles,
and shadings; this face recognition system was able to
successfully detect and authenticate most individuals'
Figure 4: Face recognition Output
faces in our dataset.
 REFERENCES
[1] NEC, “Face Detection vs Face Recognition – what’s the
difference?”,(2021).[Link]
media/face-detection-vs-facial-recognition-whats-the-difference/

[2] J. Wei, "AlexNet: The Architecture that Challenged CNNs", (2019).

[Link]

[3] Thales group, “Facial Recognition in 2021 (with examples)”, (2021).

[Link]
security/government/biometrics/facial-recognition

[4] Thales group, “Facial recognition: top 7 trends (tech, vendors, markets,
use cases & latest news)”, (2018).
[Link]
security/government/biometrics/facial-recognition

[5] The Math Works, " AlexNet convolutional neural network - MATLAB
alexnet - MathWorks Nordic", (n.d).
[Link] (2021)

[6] The Mathworks, "Series network for deep learning - MATLAB "(p. 1),
(2016).
[Link]
[Link]

[7] YouTube, “Complete Face Recognition Project Using MATLAB (Data

Collection, Model Creation And Testing)", n.d.
[Link] 2020

[8] S. Singh, S.V.A.V. Prasad, “Techniques and Challenges of Face

Recognition: A Critical Review”, Procedia Computer Science, Volume
143, (2018).

[9] M. Taskiran, N. Kahraman, C.E. Erdem, “Face recognition: Past,

present, and future (a review)”, Digital Signal Processing, Volume 106,
(2020)

[10] J.D. Woodward, Jr., C. Horn, J. Gatune, and A. Thomas, “Biometrics A

Look at Facial Recognition”,(2003).

[11] M. Gips, “How Facial Recognition Works: Everything You Need

to Know”, (2020). [Link]
works/

View publication stats