Again

REAL TIME FATIGUE MONITORING
SYSTEM WITH FACIAL EXPRESSIONS

PROJECT ID – 2025PJ-DS-20
Presented By -
Under the Supervision of -
Shruti
(2102901540105)
Shubham Rajput
(2102901540107)
Prof. KANIKA TANEJA
Tanishq Chauhan
(2102901540113)
TABLE OF
CONTENT
1. PROBLEM STATEMENT 2. OBJECTIVE 3.ABSTRACT
4. INTRODUCTION 5. LITERATURE REVIEW 6. TECHNOLOGY USED

PROBLEM
STATEMENT
•Real-time Fatigue Detection: The project aims to develop an automated,
non-invasive system using computer vision and machine learning to monitor
facial features like eye closure and yawning for detecting fatigue, providing
real-time alarms to prevent accidents in attention-demanding environments.
•Key Challenges: The system must operate efficiently in various conditions,

including changing lighting and facial structures, minimizing detection errors
(false positives/negatives) while processing live video feeds with minimal delay
for timely intervention.
OBJECTIVE
•The project aims to develop a machine learning-based system that continuously
monitors facial expressions through video streams to detect early signs of fatigue, such
as eye closure, yawning, and head position, issuing real-time warnings to prevent
accidents in high-risk environments like driving and machine operation.
• The system ensures non-intrusive monitoring with minimal user disturbance while
minimizing false alarms by leveraging robust models adaptable to varying facial
features, lighting conditions, and real-world scenarios for reliable fatigue detection.
ABSTRAC
T
•The project introduces a computer vision-based system that will track facial expressions
using a normal web camera to detect fatigue by means of metrics such as Eye Aspect
Ratio and Mouth Aspect Ratio, including symptoms like eye closure and yawning, in order
to send an alert immediately.
•The system shall provide an efficient and unobtrusive solution for the real-time detection
of fatigue with high accuracy while introducing minimal computational overhead. Target
application areas are accident prevention and increasing safety in environments that
require alertness, such as driving
INTRODUCTIO
N
•Proactive Fatigue Detection: The system analyzes key facial indicators like eye closure,
yawning, and facial muscle movements in real-time to detect early signs of drowsiness,
offering a direct and non-intrusive method for preventing accidents by focusing on the driver’s
physical state rather than erratic driving behavior.
•Road Safety Enhancement: By utilizing advanced machine learning and computer vision,
this technology aims to significantly improve road safety, reducing accidents and potentially
saving lives through timely fatigue detection.
LITERATURE
REVIEW
IoT with Firebase: Smart Ring Android App Using MAX30100 for Fatigue
Detection - May 20, 2024
Liptia Venica, Elysa Nensy Irawan, Dewi Indriati Hadi Putri (Department of Mechatronics and Artificial
Intelligence Universitas Pendidikan Indonesia Bandung)
The paper presents a Smart Ring

system utilizing the Internet of Things
(IoT) for real-time fatigue detection. It
uses the MAX30100 sensor to
measure heart rate, blood oxygen
(SpO2), and body temperature. An
Android app processes and displays
the data, achieving 100%
classification accuracy in categorizing
fatigue. The system offers real-time
monitoring and alerts for fatigue, with
minimal data processing delays.
LITERATURE
REVIEW
A Hybrid Deep Neural Network Approach to Recognize Driving
Fatigue Based on EEG Signals - June 14, 2024
Mohammed Alghanim, Hani Attar , Khosro Rezaee ,Mohamadreza Khosravi , Ahmed
Solyman ,Mohammad A. Kanan(Zarqa University, Zarqa, Jordan)
This paper introduces a hybrid deep neural

network model combining the Inception
and ResNet architectures to recognize
driver fatigue from EEG signals. It uses
spectrogram images for fatigue detection
and achieves high accuracy, with 99.1%
and 82.73% on Figshare and SEED-VIG
datasets, respectively. The proposed model
outperforms traditional methods in
classifying driver alertness and fatigue.
LITERATURE
REVIEW
Driver Fatigue Detection Using Heart Rate Variability
Features from 2-Minute Electrocardiogram Signals While
Accounting for Sex Differences - July 3, 2024
Chao Zeng,Jiliang Zhang,Yizi Su, Shuguang Li, Zhenyuan Wang, Qingkun Li and
Wenjun Wang
The paper focuses on detecting driver
fatigue using heart rate variability (HRV)
from 2-minute electrocardiogram (ECG)
signals, considering sex differences. The
study finds significant HRV changes
between alert and fatigued states for
both male and female drivers. Decision
tree models achieve high accuracy, with
sex-specific models showing superior
performance.
LITERATURE
REVIEW
Advancing driver fatigue detection in diverse lighting conditions for
assisted driving vehicles with enhanced facial recognition
technologies - May 15, 2024
Ning LinID, Yue Zuo (Nanning University, Nanning, Guangxi,
China)
The paper discusses a novel method for detecting
driver fatigue in intelligent driving assistance
systems, focusing on challenging lighting
conditions. The method integrates deep
convolutional neural networks (CNN) with
emotional state analysis to improve detection
accuracy. Validation on two datasets (Yawn and
YawDDR) shows improved accuracy, particularly
in varying lighting conditions, reaching 95.3%
accuracy on the YawDDR dataset.
LITERATURE
REVIEW
Advancements and Perspectives in Fatigue Driving Detection: A
Comprehensive Review - July 7, 2024
Xiao Yan and Ashardi bin Abas (University Pendidikan Sultan Idrts, 35900,
Malaysia)
The paper presents a comprehensive review

of fatigue driving detection methods,
including physiological signals, vehicle
dynamics, and driver behaviors. It
emphasizes multimodal information fusion
for accuracy improvement. The discussion
highlights various approaches, datasets, and
challenges, offering insights into future
advancements for effective driver fatigue
detection.
LITERATURE
REVIEW
Driver drowsiness detection system - May 26, 2024
Sachin Kumar * , Priya Devi, Meghna Singh and Meetu Rani (Department of Computer
Science and Engineering (IOT), Meerut Institute of Engineering and Technology,
Meerut)
The paper outlines a Driver Drowsiness

Detection System that uses facial recognition
and machine learning to detect drowsiness in
real-time. It leverages the dlib library to monitor
eye movements via the Eye Aspect Ratio (EAR).
When drowsiness is detected, an alarm alerts
the driver, and if unresponsive, notifications are
sent to emergency contacts.
LITERATURE
REVIEW
Fatigue detection framework using joint and facial key-point features
with dynamic distributed attention video vision transformer - June 21,
2024
Fangjie Deng Chao Yang Hui Guo Yansong Wang Lipeng Xu (Shanghai University of Engineering Sciences)
The paper presents DA-ViViT, a novel framework for

detecting driver fatigue by combining body postures
and facial features, using a Video Vision Transformer
(ViViT). The method integrates dynamic distributed
attention to enhance accuracy and reduce
computational demands, achieving 96.24% accuracy
with a significant speed improvement. The framework
effectively identifies fatigue through key body and
facial indicators, offering robust real-time monitoring
for driver safety.
LITERATURE
Facial Detection for Recognition of Drowsiness - June
REVIEW21, 2023
Abdelbari BEDJERA Khayr Eddine MESSIAID (UNIVERSITY KASDI MERBAH OF OUARGLA)
The paper focuses on detecting driver

drowsiness by facial and eye detection. It
presents a system that uses image
processing to monitor driver alertness in
real-time, particularly identifying open or
closed eyes through Support Vector
Machine (SVM) techniques. The proposed
system can handle various conditions like
lighting changes, occlusion, and glasses.
COMPARISION
Conclusion
Paper Title Objective Methodology
IoT with Firebase: Smart Ring Development of a smart ring IoT system for
Integrated MAX30100 sensor with Firebase and Android app to monitor Effective in detecting fatigue, applicable for monitoring
Android App Using MAX30100 fatigue detection.
fatigue. purposes in various domains.
for Fatigue Detection
Useful in enhancing safety, especially in driving

Facial Detection for Recognition Facial detection system for drowsiness Used embedded systems and computer vision to detect drowsiness from
environments.
of Drowsiness recognition. facial data.
A Hybrid Deep Neural Network

Hybrid neural network approach for detecting Combined CNN and RNN for continuous fatigue monitoring in drivers. Improved real-time monitoring in intelligent driving systems.
Approach to Recognize Driver
driver fatigue.
Fatigue
Driver Fatigue Detection Using

Driver fatigue detection using cognitive and
Cognitive and Biometric Sensors tracked cognitive and biometric data to identify fatigue signs. Can be integrated to alert drivers early on signs of fatigue.
biometric monitoring.
Monitoring
Advancements and
Perspectives in Fatigue Driving Review of advancements in fatigue detection for Identified future research areas in fatigue detection
Reviewed multiple studies and technologies on driver fatigue detection.
Detection: A Comprehensive drivers. technologies
Review
USE CASE DIAGRRAM
1. Regular User (Human User): The regular users will seek interaction
with the real-time monitoring system where they can access alert
history, and receive notifications on fatigue detection. In addition,
they can also access the system report summary and extract their
entire fatigue-shaped data for each session
2. Admin (Human User): The premises for Admin include keeping an
eye on the whole system carrying out tasks such as generating
alerts, sending notifications to designated contacts, and monitoring
user activity. Admin will also streamline the task of alert history
management, report generation, and certain regular system
maintenance aims in order to facilitate effective working.
3. Fatigue Monitoring System (Non-Human System): It monitors in
real time the level of user's fatigue based on facial expressions and
classifies it within specified bounds. It also provides session reports
that are detailed at some times and not detailed at other times. This
system is independent, complete, and autonomic in which human
involvement is excluded to detect real-time fatigue."
WORKFLOW
• The workflow demonstrates a comprehensive architecture for Real-Time
Fatigue Monitoring System that integrates advanced facial expression
recognition and alert mechanisms.
• The system begins by capturing live video input from the user’s device,
where it processes the feed to detect signs of drowsiness, including eye
blinks, head movement, and lip movement.
• Using facial landmark detection, the system continuously monitors the user’s
fatigue level by analysing patterns like prolonged eye closure or head tilts,
which signal fatigue. If no fatigue is detected, the system continues
monitoring the user in real time.
• Upon detecting fatigue signs, such as closed eyes for more than 10 seconds
or repeated drowsy patterns, the system autonomously triggers an alert,
starting with a local sound alarm to notify the user.
• If the drowsiness persists, the system sends an alert to designated family
members via SMS or calls using Twilio. Additionally, the system tracks and
logs session details, including fatigue incidents and responses, generating a
report at the end of each session for further analysis.
• This ensures timely intervention while maintaining comprehensive records
for ongoing fatigue management.
TECHNOLOGIES
• Face Detection Library: dlib for facial landmark detection and tracking in real-time.
• Computer Vision Libraries:
OpenCV for capturing live video feeds and processing frames for facial expression analysis.
• Video Processing Libraries: FFmpeg may still be useful for handling video files if you need to save
or manipulate the video stream.
• Real-Time Processing: Utilize threading or asynchronous programming to handle video capture

and processing concurrently for smoother performance .

Again

Uploaded by

Copyright:

Available Formats

Again

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Again

Uploaded by

Copyright:

Available Formats

REAL TIME FATIGUE MONITORING

SYSTEM WITH FACIAL EXPRESSIONS

4. INTRODUCTION 5. LITERATURE REVIEW 6. TECHNOLOGY USED

•Key Challenges: The system must operate efficiently in various conditions,

The paper presents a Smart Ring

This paper introduces a hybrid deep neural

The paper presents a comprehensive review

The paper outlines a Driver Drowsiness

The paper presents DA-ViViT, a novel framework for

The paper focuses on detecting driver

Useful in enhancing safety, especially in driving

A Hybrid Deep Neural Network

Driver Fatigue Detection Using

• Computer Vision Libraries:

• Real-Time Processing: Utilize threading or asynchronous programming to handle video capture

You might also like