Driver Attentiveness Detection Using OpenCV and Machine Learning

M. Amareswara Kumar, Shaik Sana Abida, Nayini Mounika, Yadiki Indu,

Shaik Afrin and Shaik Shahena

Department of Computer Science and Engineering, Santhiram Engineering College, Nandyal‑518501, Andhra Pradesh,

India

Keywords: Driver Monitoring, Drowsiness Detection, Computer Vision, Real‑Time Alert System.

Abstract: In today's ever-evolving world, where technology plays a crucial role, the "Driver Attentiveness Detection

System with Open CV Using Machine Learning" aims to pave the way for a safer driving experience. The

project uses Open CV based computer vision techniques and machine learning algorithms to identify early

signs of driver drowsiness and distraction. The intelligent monitoring system can analyze and determine

whether the driver is looking at the road or yawning and communicate any potential out-of-control situation

to the driver using facial features and indicating different aspects like the position of (head, eyes) and blinking

rate/jaw yawning frequency. The advanced model is fed with real-time video feed from a camera placed inside

the vehicle and has been programmed to recognize facial landmarks to identify alert and drowsy states. Upon

detection of drowsiness or distraction, it immediately conjures real-time alerts via alarms or notifications,

redirecting the driver’s focus. This project addresses this issue with the purpose of reducing road accidents,

which are one of the leading causes of human deaths and injury. Automated monitoring and AI-driven

decision-making provide a strong solution for driver protection, especially for long distance drivers, fleet

drivers and self-driving vehicles.

1 INTRODUCTION

“Driver Attentiveness Detection System with Open

CV Using Machine Learning” The above concept of

driver behavior monitoring system for real time

monitoring of driver behavior will help to become a

big breakthrough in traffic safety (Charbonnier et al.,

2008; Borghini et al., 2012). Road Condition

Monitoring system is used inside the vehicle so as not

to lose the alertness of drivers so that they can react

well and proper while driving because the cause of

most traffic accidents is because of fatigue and

distraction (Jap et al., 2009; Abdul Rahmat et al.,

2012). Using computer vision and machine learning,

this initiative also fills a vital gap in the area of

transportation safety (Chaitanya et al., 2024).

This technology is then utilizing a camera which

is being placed at a prime location in the car which

transmits live video clips of the driver for processing.

The system uses algorithms that detect facial

landmarks to examine features such as the position of

the head, the frequency of blinking and eye

movement (Charbonnier et al., 2008). By tracking

these cues, the system is able to distinguish between

when one is in alert versus inattentive phases. When

the driver becomes inattentive or sleepy, the system

triggers real-time alarms urging the driver to pay

attention and keep his/her eyes on the road (Borghini

et al., 2012).

Open CV is a library of programming functions

used for real-time computer vision, helps analyze

video streams and tracking the driver actions over the

time. Behavior classification has been performed

using multiple machine learning algorithms

specifically with the help of Convolutional Neural

Network (CNN's) methods which efficiently

recognize the signs of drowsiness and distraction (Jap

et al., 2009; Devi et al., 2023). The emotional state

and recommendation level of the driver is analyzed

by facial recognition algorithms. These technologies

create a complete framework for the improvement of

driver safety, as well as the capacity for airlines and

other long-distance fleets to create intelligent,

responsive vehicles (Parumanchala Bhaskar et al.,

2024).

640

Kumar, M. A., Abida, S. S., Mounika, N., Indu, Y., Afrin, S. and Shahena, S.

Driver Attentiveness Detection Using OpenCV and Machine Learning.

DOI: 10.5220/0013870600004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 1, pages

640-644

ISBN: 978-989-758-777-1

2 LITERATURE REVIEW

2.1 Machine Learning

An area within AI, machine learning (ML) is the field

of study that gives computers the ability to learn

without being explicitly programmed. In

conventional programming, all instructions for

executing a task are given explicitly to the program.

These processes fall into one of three categories:

supervised learning training models on labeled data;

unsupervised learning finding patterns in unlabeled

data; or reinforcement learning motivating agents to

learn through trial and error. Jacobs University, for

instance, has adopted machine learning with a

vengeance, applying it to predictive analytic, fraud

detection, recommend systems, speech recognition,

and other areas. Which makes it specifically

beneficial for industries like cyber security, health

care or any form of automation aimed at better

decision making based on the real-time processing

and analysis of extremely large data sets (Mahammad

et al., 2024; Sunar & Viswanatham, 2018).

2.2 The Use of Open CV to Promote

Attentiveness

Open CV also monitors facial features from a video

feed from the driver to calculate driver attentiveness.

To assess if a driver is distracted, it watches for key

indicators such as head position, mouth openings and

eye blinks (Charbonnier et al., 2008; Devi et al.,

2023). This enables timelier alerts that increase

safety.

Open CV tracks how far away the driver’s head

turns from the road, and can send alerts to the driver

to re-engage their concentration. The company's

diligent video analysis makes certain that any signs

of inattention or drowsiness are instantly detected,

which is crucial in preventing accidents (Borghini et

al., 2012).

Accuracy can also be improved using Open CV

along with machine learning algorithms. With

knowledge drawn from such a large pool of data, the

system can identify various degrees of driver

attention and cater to personalities. If it detects

distraction, open CV can also provide alerts sounds or

visual signals on the dashboard to remind the driver

to pay attention (Jap et al., 2009; Paradesi Subba Rao,

2024).

2.3 The Use of Machine Learning to

Promote Driver Attentiveness

Machine learning analyzes data from cameras to

assess where the drivers are looking, and what they

are doing with their phones, to monitor their attention

(Charbonnier et al., 2008; Chaitanya et al., 2024).

That visual data is then run through algorithms, in

this case, traditional neural networks (CNN’s)

searching for signs of distraction, like looking at a

phone or generally turning your head away from the

road (Devi et al., 2023).

2.3.1 Analysis of Head Movements

The system tracks the position of the head to see

whether the driver gaze has been diverted, in other

words, not looking. It determines whether the driver’s

attention is on the road or elsewhere by monitoring

angles and head positions. The algorithms can also

detect when a driver is using a mobile phone by

recognizing certain movements, reaching for the

phone and looking down at it (Borghini et al., 2012).

2.3.2 Real-Time Feedback

These technologies combine to offer monitoring in

real-time. The driver may receive alerts reminding

them to keep their eyes on the road if the system

detects distraction (Parumanchala Bhaskar et al.,

2024).

2.3.3 Data Collection and Training

To increase precision, the system is trained using

sizable image collections that depict drivers in

various attentiveness levels. The models improve

over time by learning from

fresh data (Mr. M.

Amareswara Kumar, 2024; Meem, 2023).

3 METHODOLOGY

This section describes the methodology used to detect

the driver attentiveness; in particular, section 3.1

describes the project architecture, section 3.2

describes the data set information, section 3.3

describes the feature extraction approach.

3.1 Architecture

Figure 1 illustrates the System Architecture.

Driver Attentiveness Detection Using OpenCV and Machine Learning

641

Figure 1: System Architecture.

3.2 Dataset Information

Video Data: A set of videos that show different

driving situations that show both attentive and

inattentive driving.

Labels of Ground Truth: Examples of attentive

driving include situations in which the motorist is

closely observing the road.

Distracted (0): Situations in which the motorist is

distracted by something, such as looking away or

using a cell phone.

3.3 Feature Extraction

• F1: Eye Attention: Open and closed eye

states are analyzed to assess attentiveness;

loss of concentration or closed eyes indicate

distraction.

• F2: Mouth Movement: Tracking mouth

movements to identify chatting or other

behaviors that might be signs of a phone-

related distraction.

• F3: Head Positioning: Examine the tilt and

direction of the head; notable departures

from a forward-facing posture may suggest

preoccupation.

• F4: Temporal Analysis: Assesses the length

and frequency of interruptions to gain

insight into trends in inattention over time.

• F5: Driver State Classification: For machine

learning purposes, categorize driving

behavior occurrences into "attentive" (1) or

"distracted" (0) categories throughout the

model training phase.

The system seeks to efficiently monitor driver

attentiveness and eventually improve road safety by

adhering to this standardized technique.

4 IMPLEMENTATION AND

RESULTS

In this section the implementation details are

mentioned to detect the jamming attacks. 4.1 Section

contains the model selection and 4.2 section contains

the results of the implements.

4.1 Model Selection

4.1.1 Supervised Learning Algorithms

Model 1: CNNs (convolutional neural networks).

CNNs are perfect for analyzing the eye and facial

states of drivers because they work especially well

with image data. Pooling layers lower dimensionality,

convolutional layers learn spatial hierarchies of

features, and fully connected layers categorize the

driver's attentiveness. Real-time processing is a

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

642

strength of CNNs, which enables prompt feedback on

driver conditions.

Model 2: SVM (support vector machine).

SVM uses features like head posture and eye aspect

ratio that are taken from images to distinguish

between alert and sleepy states. Classifying intricate

patterns in driver behavior can be accomplished with

SVM because it works well in high-dimensional

spaces, despite its potential for slower data

processing.

Model 3: Random Forests.

Random Forests use a collection of decision trees to

improve classification accuracy. The XGBoost model

is resistant to overfitting and deals efficiently with

diverse features of driver attentiveness such as head

positioning and eye openness by averaging multiple

trees.

Model 4: K-Nearest Neighbors (KNN).

KNN classifies the driver's status based on the

similarity of its behavior with the closest data points.

If most of the surrounding points are attentive, the

new observation is classified as attentive, while if

they are drowsy, the new observation is classified as

drowsy. KNN will rapidly classify data based on the

distance to existing data points in real-world

scenarios and is simple to explain and implement.

Model 5: Decision Tree Classifier.

In a basic model, the Decision Tree classifier can be

very simple and intuitive for classification. This

allows one to visualize the decision-making process

behind driver attentiveness in an easy manner by

splitting the data into branches based on specific

features.

The sixth model is Multi-Layer Perceptron (MLP).

4.1.2 Unsupervised Learning Algorithms

Unsupervised learning algorithms, which do not

require labelled data, are primarily used for

clustering and anomaly detection in driver

attentiveness detection.

Model 7: K-Means Clustering.

K-Means clustering can be used to help classify

different levels of driver attentiveness using feature

vectors extracted from facial images. By using a

clustering algorithm this enables us to hear the

signals of frequent behaviors in the drivers such as

awake, drowsy, distracted, etc.

PCA can be useful for visualizing and

understanding high-dimensional datasets after the

dimensionality reduction and feature extraction. In

the case of driver attentiveness detection, PCA can

reduce the size of the feature space, which can aid in

the identification of significant patterns and links

between the photo-derived features.

4.2 Results

Figure 2: Roc-Auc Scores of Test Data.

ROC-AUC curve (figure 2) is a graphical

representation of the true positive rate against the

false positive rate. The ROC-AUC curve represents

the true positive rate against the false positive rate at

various threshold settings and was implemented to

test the models. Performance: The models'

effectiveness in predicting driver inattention varied,

but Random Forest and Support Vector machine

yielded the highest accuracies according to the

aforementioned metrics. This systematic study

illustrates the multiple ML models as well as

potential applications involved in driver attention

monitoring. Each model performs a different function

when it comes to assessing head, eye and mobile

phone movements during driving.

Driver Attentiveness Detection Using OpenCV and Machine Learning

643

5 CONCLUSIONS

Here in this research, we proposed a machine learning

model and OpenCV solution to detect the driver

attention. Models that performed best at connecting

with driver inattention were classified as Random

Forest and Support Vector Machine classifiers. We

perform classification tasks using performance

measures on these features, which will gather key

face attributes such as yawning, head tilt, and eye

gazes to reflect how well our methodology can

differentiate between attentive and distracted states.

This real-time monitoring system provides timely

alerts, greatly enhancing road safety and is useful for

fleet management, autonomous vehicles and long-

distance drivers.

REFERENCES

A. Abdul Rahmat, M. Siti Atiqah, L. Fauziana, and Z.

Ahmad Noor Syukri, "MIROS crash investigation and

reconstruction: annual statistical 2007-2010," 2012.

B. T. Jap, S. Lal, P. Fischer, and E. Bekiaris. "Using EEG

spectral components to assess algorithms for detecting

fatigue," Expert Systems with Applications, vol. 36, pp.

2352-2359, 2009.

Chaitanya, V. Lakshmi, et al. "Identification of traffic sign

boards and voice assistance system for driving." AIP

Conference Proceedings. Vol. 3028. No. 1. AIP

Publishing, 2024

Devi, M. Sharmila, et al. "Extracting and Analyzing

Features in Natural Language Processing for Deep

Learning with English Language." Journal of Research

Publication and Reviews 4.4 (2023): 497-502

G. Borghini, L. Astolfi, G. Vecchiato, D. Mattia, and F.

Babiloni, "Measuring neurophysiological signals in

aircraft pilots and car drivers for the assessment of

mental workload, fatigue, and drowsiness,"

Neuroscience & Biobehavioral Reviews, 2012.

Mahammad, Farooq Sunar, et al. "Key distribution scheme

for preventing key reinstallation attack in wireless

networks." AIP Conference Proceedings. Vol. 3028.

No. 1. AIP Publishing, 2024.

Mr.M.Amareswara Kumar, "Baby care warning system

based on IoT and GSM to prevent leaving a child in a

parked car" in International Conference on Emerging

Trends in Electronics and Communication Engineerin-

g-2023, API Proceedings July-2024.

Mr.M.Amareswara Kumar, Effective Feature Engineering

Technique For Heart Disease Prediction With Machine

Learning" in International Journal of Engineering &

Science Research, Volume 14, Issue 2, April-2024 with

ISSN 2277-2685.

Paradesi Subba Rao, "Detecting malicious Twitter bots

using machine learning" AIP Conf. Proc. 3028, 020073

(2024), https://doi.org/10.1063/5.0212693

Parumanchala Bhaskar, et al. "Incorporating Deep Learning

Techniques to Estimate the Damage of Cars during the

Accidents" AIP Conference Proceedings.

S. Charbonnier, A. Picot, and A. Caplier, "On-line

automatic detection of driver drowsiness using a single

electroencephalographic channel," in Engineering in

Medicine and Biology Society, 2008. IEEE 30th

Annual International Conference on Business and

Management, 2008, pp. 3864–3867.

Sunar, Mahammad Farooq, and V. Madhu Viswanatham.

"A fast approach to encrypt and decrypt of video

streams for secure channel transmission." World

Review of Science, Technology and Sustainable

Development 14.1 (2018): 11-28.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

644