Project Report

on
Gaze Tracker

Submitted in the partial fulfillment of the requirements for

the award of Degree of B. Tech

By
Khushi (2100680100186)
Kunal Yadav (2100680100195)
Manav Adhana (2100680100204)
Maneesha (2100680100205)

Under the Supervision of: -

Dr. Satendra Kumar
(Associate Professor, Department of CSE)

Department of Computer Science & Engineering

Meerut Institute of Engineering and Technology,
Meerut – 250005

Dr. A.P.J. Abdul Kalam Technical University, U.P., Lucknow

[Batch: 2021-2025]
 DECLARATION

I hereby declare that this submission is my own work and that, to the best of my knowledge
and belief, it contains no material previously published or written by another person nor
material which to a substantial extent has been accepted for the award of any other degree
or diploma of the university or other institute of higher learning, except where due
acknowledgment has been made in the text.

Signature:

Name : Khushi

Roll No.: 2100680100186

Date :

Signature:

Name : Kunal Yadav

Roll No.: 2100680100195

Date :

Signature:

Name : Manav Adhana

Roll No.: 2100680100204

Date :

Signature:

Name : Maneesha

Roll No.: 2100680100205

Date :

I
 CERTIFICATE

This is to certify that Project Report entitled ― GAZE TRACKER which is submitted
by Kunal Yadav (2100680100195), Manav Adhana (2100680100204), Maneesha
(2100680100205) and Khushi (2100680100186) in partial fulfillment of the requirement
for the award of degree B. Tech. in Department of Computer Science and Engineering of
Dr. A.P.J. Abdul Kalam Technical University, U.P., Lucknow., is a record of the
candidate own work carried out by him/her under my/our supervision. The matter
embodied in this Project report is original and has not been submitted for the award of
any other degree.

Date: Supervisor

Dr. Satendra Kumar

(Associate Professor, CSE)

II
 ACKNOWLEDGEMENTS

It gives us a great sense of pleasure to present the report of the B. Tech Project undertaken
during [Link]. Final Year. We owe special debt of gratitude to our guide Dr. Satendra Kumar,
Department of Computer Science & Engineering, Meerut Institute of Engineering and
Technology, Meerut for his constant support and guidance throughout the course of our work.
His sincerity, thoroughness and perseverance have been a constant source of inspiration for us.
It is only his cognizant efforts that our endeavors have seen light of the day.
We also do not like to miss the opportunity to acknowledge the contribution of all faculty
members of the department for their kind assistance and cooperation during the development
of our project. Last but not the least, we acknowledge our friends for their contribution in the
completion of the project.

Signature:

Name : Khushi

Roll No. : 2100680100186

Date :

Signature:

Name : Kunal Yadav

Roll No. : 2100680100195

Date :

Signature:

Name : Manav Adhana

Roll No. : 2100680100204

Date :

Signature:

Name : Maneesha

Roll No. : 2100680100205

Date :

III
 ABSTRACT

This research paper describes the implementation of Raspberry Pi technology to build a head-
movement and blink-controlled PC system for individuals with physical disabilities. By
recognizing head movements and eye blinks, the system allows users to control the computer
cursor without the need for traditional input devices. The hardware setup includes a Raspberry Pi
3B for processing, a 720p USB camera to capture video of the user's face, and several infrared
(IR) sensors that detect eye blinks. The camera tracks head movements to move the cursor in
different directions, while the IR sensors detect intentional blinks that simulate mouse clicks.

The system is built using the Python programming language and Open CV, a powerful computer
vision library, to handle face detection and cursor mapping. Through a combination of hardware
interfacing and real-time software control, the system provides a smooth and responsive user
experience. A key part of the implementation involves calibration and testing to ensure that the
system accurately detects user input and avoids errors caused by unintentional movements or
blinks.

The ability to differentiate between reflexive and intentional blinks, using IR sensors and
debounce logic, significantly enhances the reliability of cursor control. As a result, users can
interact with a computer in a more natural and independent manner. This approach highlights
how technology can be adapted to serve accessibility needs by focusing on user-friendly and cost-
effective solutions.

Looking forward, this system can be improved with more refined blink detection techniques,
enhanced camera quality for better head or eye tracking, and additional input methods such as
gesture recognition or voice activation. These improvements could further increase usability and
extend the system's application to mobile devices and smart home controls. The core goal remains
to provide physically challenged individuals with a greater degree of freedom and interaction
with digital technology, without relying on complex or expensive tools.

IV
 TABLE OF CONTENTS

TITLE ........................................................................................................ Page No.

Declaration………………………………………………………………………………….………...i
Certificate……………………………………….…………...……………………...………. ii
Acknowledgement………………………………………………………………………...…iii
Abstract……………………………………………………………….…………………......iv
List of Table……………………………………………………………………...................vii
List of Figures……………………………………………………………………..……….viii

CHAPTER 1 – INTRODUCTION .......................................................................................... 1

1.1 Overview ............................................................................................................ 1
1.2 Feature Extraction .............................................................................................. 2
1.3 Accessing Models .............................................................................................. 2
1.4 Summary of Result............................................................................................. 3
1.5 Scope................................................................................................................... 4
1.6 Software development methodology ................................................................. 5
CHAPTER 2 –LITERATURE REVIEW ............................................................................. 7
2.1 Key Factors Influencing .....................................................................................7
2.2 Strategies ............................................................................................................8
CHAPTER 3 – SOFTWARE REQUIREMENTS SPECIFICATIONS ........................... 10
3.1 Introduction. ....................................................................................................... 10

3.2 General architecture of software ....................................................................... 10

3.3 Requirements Specifications. ............................................................................ 11

3.3.1 Functional Requirements............................................................................... 11

3.3.2 Non-Functional Requirements. ..................................................................... 12
3.4 System Requirement Study ............................................................................. 13

V
 3.4.1 Software Requirements .................................................................................. 13
3.4.2 Hardware Requirements. ............................................................................... 14
3.5 System Design………………………………………………………………...15

3.5.1 Introduction…………………………………………………………………15
CHAPTER 4–IMPLEMENTATION .............................................................................. 18
4.1.1 Code Implementation…………………………………………………………...19

CHAPTER 5– TECHNOLOGY USED........................................................................... 21

CHAPTER 6- RESULT AND DISCUSSION .................................................................22

CHAPTER 7- CONCLUSION & FUTURE WORK .....................................................23

REFERENCES .................................................................................................................. 25

Appendix.............................................................................................................................27

Research Paper1................................................................................................................31

Research Paper2................................................................................................................37

Project Plagiarism Report................................................................................................53

VI
 LIST OF TABLES

Table No. Description Page no.

1.1 Expected Accuracy 4

1.2 Features 4

6.1 Testing 22

6.2 Optimal Condition 22

VII
 LIST OF FIGURES

Fig. No. Description Page no.

3.1 General Architecture 11

3.2 Non-functional Requirement 12

3.3 System Requirement 13

3.4 System Design Diagram 17

4.1 Raspberry Pi setup 18

4.2 Wearable assistive device 18

4.3 Smart glasses prototype 19

4.4 Code Implemented 19

4.5 Code Implemented 20

4.6 Code Implemented 20

VIII
 CHAPTER 1
INTRODUCTION

1.1 OVERVIEW

Introduction to Head and Blink Tracking

Head and blink tracking is a technique used to monitor and analyse head movements and eye
blinks to control computer interfaces, especially for individuals with physical disabilities. It has
wide applications in assistive technologies, human-computer interaction (HCI), and
accessibility solutions. By tracking head movements and detecting intentional blinks, such
systems allow users to control the computer cursor and clicks without traditional input devices.
A head and blink tracking system typically consists of a camera that captures real-time video
of the user's face and infrared (IR) sensors that detect eye blinks. The system processes these
signals to detect facial features (such as eyes, nose, and mouth) and uses these to estimate head
orientation and blink events. This technology enables hands-free control of devices, enhancing
accessibility for individuals with mobility impairments and improving user experience in
assistive applications.
Unlike AI-based gaze tracking systems that require complex machine learning models,
traditional head and blink tracking relies on lightweight image processing and sensor-based
techniques. These methods are efficient, cost-effective, and suitable for real-time applications
with minimal hardware requirements.

How Head and Blink Tracking Works

1. Face Detection & Head Tracking: A USB camera captures real-time images of the
user's face. Facial landmarks (eyes, nose, mouth) are detected, and head movement is
mapped to cursor direction.
2. Blink Detection: IR sensors mounted near the eyes detect intentional blinks by sensing
quick proximity changes.
3. Cursor & Click Control: Head movements control cursor direction, while detected
blinks trigger click actions.
4. Calibration & Real-time Processing: The system undergoes calibration to fine-tune
sensitivity and continuously updates cursor position and click detection in real time.

Applications of Gaze Tracking

• Assistive Technology: Enables individuals with physical disabilities to control

computers and smart devices using only their eye movements.

• Human-Computer Interaction (HCI): Supports hands-free navigation in computers and

interactive systems.

• Virtual Reality: Enhances user experience by enabling gaze-based interactions.

1
 • Medical Research: Used in studies related to vision disorders, cognitive conditions, and
neurological diseases.

• Marketing & Advertising: Helps analyze user engagement by tracking eye movement
on advertisements and digital content.

• Automotive Safety: Detects driver fatigue and distraction, reducing accident risks by
improving road safety.

• Smart Home Controls: Facilitates hands-free control of home devices for users with
limited mobility.

1.2 Feature Extraction

Feature Extraction in Gaze Tracking

Feature Extraction in Head and Blink Tracking

Feature extraction in this system involves identifying key face-related and eye-related elements
from camera images and sensor signals. The main features extracted include:
• Facial Landmark Detection: Using models like dlib or OpenCV, the system detects
key points on the face such as eyes, nose, and mouth for head orientation tracking.
• Head Pose Estimation: By calculating angles based on landmark positions, the system
estimates head direction (up, down, left, right) to move the cursor.
• Blink Detection (via IR Sensors): Infrared sensors detect intentional blinks by
measuring quick changes in IR reflection when eyes close, differentiating voluntary
blinks from natural ones.
• Cursor Mapping: Detected head movements and blinks are mapped to cursor
movement and mouse click actions on the computer screen.

Implementation in Gaze Tracking:

1. Data Collection: Camera captures face video; IR sensors capture blink signals.
2. Feature Extraction: Detect face landmarks for head tracking and sensor signals for blink
detection.
3. Movement & Click Estimation: Head pose estimation algorithms map movements to
cursor control; blink signals trigger click actions.
4. Real-Time Processing: System continuously processes camera and sensor data to
provide smooth, responsive control.

1.3 Accessing Models

This system uses traditional image processing and sensor-based methods rather than AI models,
making it lightweight and efficient.
Key Tools and Methods:

2
 • OpenCV Library: Used for face detection, facial landmark tracking, and image
processing.
• dlib Library: Provides accurate facial landmark detection models for real-time head pose
estimation.
• IR Sensors and GPIO Programming: Hardware-based blink detection using infrared
proximity sensors connected to Raspberry Pi GPIO pins.
• Geometric Models: Trigonometric calculations to convert head tilt angles into screen
coordinates.
• Custom Algorithms: Rule-based detection and calibration methods for filtering out false
positives and tuning sensitivity.

1.4 Summary of Results:

This project implements a head-movement and blink-based control system, providing an

alternative interaction method for individuals with physical disabilities. Instead of using
full gaze tracking, the system relies on head movement detection and intentional eye
blinks to simulate cursor control and clicking. This method avoids complex eye tracking
while still offering reliable hands-free operation.

Unlike AI-based gaze tracking, which relies on machine learning to predict gaze
direction, this system uses traditional image processing and sensor-based techniques.
Head movements are detected by tracking facial landmarks, while blinks are registered
through infrared (IR) sensors that capture changes around the eye area.
The system's performance is evaluated based on the following criteria:

1. Head Movement Detection Accuracy – The ability to correctly identify

directional head tilts using facial landmark tracking and map them to cursor
movement.
2. Blink Detection Reliability – The system’s effectiveness in distinguishing
intentional blinks from reflexive ones using IR sensor input and threshold
filtering.
3. Processing Speed – The responsiveness of the system, ensuring real-time control
with minimal delay during operation.

The accuracy and responsiveness depend on several factors, including camera resolution,
lighting conditions, and sensor placement. Typically, for systems built with a Raspberry
Pi 3B, a 720p USB camera, and IR sensors, the overall accuracy ranges from 85% to
92% in controlled indoor environments. Proper lighting, stable head positioning, and
careful calibration steps are essential for achieving peak performance.

By combining geometric face models and carefully analyzing head position and blink
patterns, the system achieves dependable control of cursor movement and clicks. While
occasional discrepancies may occur due to rapid head movements, varying facial
structures, or light interference, regular recalibration and optimized sensor alignment
help maintain consistent performance. This level of accuracy is sufficient for assistive

3
applications such as menu navigation and basic computer interaction tasks.

Table:1.1 Expected Accuracy

Table:1.2 Features

1.5 SCOPE

Scope of Head and Blink Tracking System:

Head and blink tracking technology offers an innovative and accessible method for individuals,
especially those with physical and motor impairments, to control digital devices through natural
movements. By leveraging simple head tilts and intentional eye blinks, users are empowered to
perform complex tasks such as cursor movement and click actions without relying on traditional
input devices like a mouse or keyboard. This makes it particularly valuable in creating inclusive
technologies that bridge the gap between humans and machines.
In the field of assistive technology, this system plays a transformative role. It provides a lifeline
for individuals living with conditions such as Amyotrophic Lateral Sclerosis (ALS), cerebral
palsy, paralysis, spinal cord injuries, and muscular dystrophy. By enabling cursor control and
click simulation solely through head movements and blink detection, the technology restores
independence and dignity to users, allowing them to navigate software applications, browse the
internet, and communicate with others in a hands-free manner.
In the domain of gaming and virtual reality (VR), head tracking enhances immersive
experiences by allowing players to naturally interact with virtual environments. Users can aim,
navigate, and make selections simply by moving their heads, providing a more intuitive and
engaging gameplay experience. Similarly, in augmented reality (AR), combining head
gestures with blink-based selections can further improve user engagement, offering a seamless
interaction model where digital elements respond instantly to physical cues.
The automotive industry also stands to benefit from this technology. Blink detection and head
movement monitoring are increasingly being integrated into advanced driver assistance systems
(ADAS) to detect signs of driver fatigue, drowsiness, and distraction. By issuing alerts when
drivers show signs of inattention or microsleeps, these systems enhance road safety, reduce

4
accident rates, and contribute to a safer driving environment for everyone.
In medical research and healthcare, analyzing blink rates and head movement patterns
provides critical insights into neurological and cognitive conditions. Researchers and clinicians
can use this technology to study disorders such as Parkinson's disease, Alzheimer's disease,
autism spectrum disorders, and Attention Deficit Hyperactivity Disorder (ADHD). This opens
up possibilities for non-invasive diagnostics and continuous patient monitoring, enabling early
detection and intervention.
The field of smart home automation presents another exciting application. As smart
environments evolve, integrating head and blink tracking systems can enable users to control
lighting, appliances, entertainment systems, and security devices through simple, hands-free
commands. For individuals with limited mobility, this technology offers greater independence,
enabling them to interact with their living space in a seamless and intuitive manner.
With the rise of wearable technologies, such as smart glasses and head-mounted displays, head
and blink tracking systems are being adapted for professional and industrial use. Surgeons in
sterile operating rooms can navigate medical imaging systems using only eye gestures,
eliminating the need for physical contact. Similarly, engineers, field technicians, and
emergency responders can access information, control tools, and communicate while keeping
their hands free for critical tasks, improving efficiency and safety in demanding environments.
Furthermore, as hardware miniaturization and sensor precision continue to advance, the
potential applications of this technology will expand even further. Emerging fields such as
robotics, human-computer interaction (HCI), and wearable health monitoring can benefit
from reliable, non-invasive head and blink tracking systems. Future iterations may include
gesture recognition, voice control integration, and real-time adaptive interfaces, making digital
systems more responsive and accessible to a wider audience.

In conclusion, while the current project primarily focuses on providing cursor control and click
simulation for individuals with physical disabilities, the underlying head and blink tracking
technology has a broad and evolving scope.

1.6 SOFTWARE DEVELOPMENT METHODOLOGY

The software development process for the gaze tracking system follows an Agile methodology,
which emphasizes iterative progress, continuous testing, and user feedback. Agile allows the
development team to adapt quickly to changing requirements and ensure a high-quality product.

Each development cycle (or sprint) focuses on implementing specific features such as blink
detection, head movement tracking, cursor control, or calibration routines. At the end of each
cycle, the software is tested in real-time to validate its performance and identify areas for
refinement. This incremental approach not only enhances system stability and reliability but
also ensures that user needs are consistently prioritized during development.

Regular user testing plays a crucial role in evaluating the system’s usability, responsiveness,
and accuracy. Feedback from individuals with physical impairments helps the developers fine-
tune features like blink sensitivity, cursor speed, and dwell time for selections. This close loop
of feedback and adjustment ensures the final product is intuitive, accessible, and practical for
real-world use.

Moreover, Agile's flexibility allows the team to explore improvements in user interface design
and system calibration without overhauling the entire application. It also encourages
collaboration between software engineers, hardware integrators, and user experience designers,

5
resulting in a well-rounded, user-friendly solution.

In the future, this development approach makes it easier to introduce additional features—such
as voice commands or gesture-based controls—and to adapt the system for use on other
platforms, such as tablets, smartphones, or home automation devices. The focus remains on
building a reliable, scalable, and maintainable system that can improve digital accessibility for
users with mobility limitations.

Key Stages of Software Development

• Requirement Analysis: Understanding user needs and defining system functionalities,

such as real-time gaze tracking, calibration, and data visualization.

• Design Phase: Creating system architecture, selecting appropriate image processing

techniques, and designing user interfaces for gaze-based interaction.

• Implementation: Developing the software using programming languages like Python ,

along with libraries such as OpenCV for image processing. Ensuring modular and
scalable code.

• Testing: Conducting unit tests, integration tests, and performance tests to validate
accuracy, response time, and user experience.

• Deployment & Maintenance: Deploying the gaze tracker on target platforms,

monitoring performance, and providing updates based on user feedback.

6
 CHAPTER 2
LITERATURE REVIEW

Gaze Tracking Technology

Head movement and blink detection technology are key methods used for hands-free computer
control, especially for individuals with physical disabilities. These systems track the movement
of a user’s head to move the mouse cursor and detect intentional eye blinks to simulate mouse
clicks. The technology is often implemented with a combination of a camera and infrared (IR)
sensors. The camera captures video of the user's face to track head movements, while IR sensors
detect the user’s blinks, allowing for click actions without using a traditional input device like
a mouse.
Unlike traditional gaze tracking, which focuses on eye movement and pupil detection, the
system in our project tracks head position through face detection techniques and interprets eye
blinks for click actions. This approach provides a straightforward, cost-effective solution that
can be deployed using devices like Raspberry Pi and USB cameras. It offers significant value
for users with mobility impairments, enabling them to interact with digital interfaces by simply
moving their heads and blinking.
The system is relatively simple, yet effective. It does not require complex or expensive tools,
and with the right calibration and real-time software control, it can provide reliable and smooth
interaction with the computer. The use of IR sensors allows for accurate blink detection, which
reduces errors from reflexive blinks and ensures that only intentional blinks are considered as
click actions.

2.1 Key Factors Influencing Gaze Tracking

Several key factors influence the accuracy, efficiency, and adoption of gaze tracking systems.
These factors impact the reliability of gaze tracking techniques and their practical usability
across different domains.

• Lighting Conditions: Variations in lighting can significantly affect the visibility of eye
features. Low-light environments or excessive glare can interfere with pupil detection,
reducing the accuracy of gaze tracking systems.

• User Variability: Differences in individual eye structures, blinking frequency, and head
movements present challenges in ensuring consistent gaze tracking performance across
diverse users.

• Camera Quality and Resolution: Higher-resolution cameras provide more accurate gaze
tracking but require greater computational power. The frame rate of cameras also affects
real-time tracking performance.

• Real-time Processing Requirements: Applications such as gaming, assistive technology,

and virtual reality demand low-latency gaze tracking for seamless interaction. Efficient
image processing algorithms.

7
2.2 Strategies

Accurate and Reliable Head Movement Tracking

• Use traditional face detection algorithms like Haar cascades (as used in OpenCV) to
detect and track head position.
• Implement image processing techniques to map head movements to screen
coordinates, ensuring smooth cursor movement in real-time.
• Calibrate the system to account for variations in user head positions and movement
speeds.
Effective Blink Detection
• Use infrared sensors to accurately detect eye blinks by identifying changes in light
levels near the eyes.
• Implement debounce mechanisms to prevent false positives caused by reflexive blinks.
• Fine-tune blink detection sensitivity to distinguish between unintentional and
intentional blinks.
Seamless User Experience
• Ensure low-latency processing to maintain a responsive system that updates cursor
positions in real-time.
• Design adjustable sensitivity settings for both head movement tracking and blink
detection, allowing users to fine-tune the system to their preferences and reduce false
detection.
• Provide intuitive interface designs that make interaction easy without requiring
extensive training or learning curves.
Personalization Without AI
• Use rule-based systems for head tracking and blink detection that rely on pre-
programmed thresholds and algorithms.
• Allow manual customization for settings like blink dwell time or head movement
sensitivity to adapt the system to individual user needs.
• Enable adjustable click options, where a user can configure how long to hold the blink
to register a click.
Head Movement and Blink-Based Navigation
• Implement head gestures (e.g., tilting the head left/right or up/down) to enable
navigation through menus or applications.
• Use blinks as a trigger for selection, allowing users to open files, click on buttons, or
perform other mouse-related actions.
• Enable multi-screen support for users to control multiple displays with ease.
Customer Support and Engagement
• Provide on-screen tutorials that demonstrate how to use the head movement and blink-
based interface, guiding users through the system’s features.
• Set up help menus that allow users to troubleshoot issues using head movements and
blinks.
• Implement feedback mechanisms, where users can report any issues or difficulties they
encounter while interacting with the system.
Data Security and Privacy Protection
• Ensure that any head movement data or blink logs are stored securely and are only
used for system calibration and improvement.
• Allow users to control permissions regarding the use of their data, ensuring that it is

8
 only collected when necessary for system functionality.
• Encrypt all sensitive interaction logs and respect user privacy, especially in compliance
with relevant privacy laws.
Reward-Based Engagement Strategies
• Develop reward systems that encourage users to complete certain tasks using head
movement and blink detection, such as interactive tutorials or loyalty programs.
• Create gamified experiences where users can earn rewards by completing head-
controlled challenges.
• Offer discounts or rewards to users who consistently engage with the system,
motivating long-term use.

Several prior research efforts have explored assistive technologies leveraging eye and head
movements for computer interaction, particularly targeting individuals with physical
disabilities. For instance, Álvarez et al. (2013) presented an eye-gaze tracking system for
human-computer interaction using low-cost webcams, demonstrating promising accuracy for
basic pointer control. Similarly, Villanueva and Cabeza (2007) emphasized vision-based
systems for eye tracking that could be implemented in real-time environments using
commercial-grade cameras. More recent developments such as the Eye Writer project (Golan
et al., 2010) employed open-source hardware and software to enable paralyzed individuals to
draw and communicate using eye movements. Furthermore, systems utilizing blink detection
for mouse click simulation were studied by Maazaoui et al. (2018), who integrated infrared
sensors with a neural network classifier to distinguish intentional blinks. These related studies
have laid the foundation for accessible, affordable, and real-time assistive technologies,
aligning well with the proposed Raspberry Pi-based system using Python and OpenCV. The
current work builds on these advancements by incorporating head tilt and blink-based controls
using minimal hardware, which contributes to the growing field of hands-free interfaces for
physically challenged users.

9
 CHAPTER 3
SOFTWARE REQUIREMENTS SPECIFICATION

3.1 INTRODUCTION

This chapter outlines the software requirements for the head and blink movement tracker system.
It defines the system architecture, functional and non-functional requirements, feasibility study,
and hardware/software requirements necessary for successful implementation.

The system architecture describes the interaction between key components, such as the Raspberry
Pi, USB camera, and infrared sensors. The software processes real-time video input to detect head
tilts and eye blinks, which are translated into cursor movement and click commands. The
architecture follows a modular design to ensure clarity, ease of development, and future scalability.

Functional requirements specify the essential tasks the system must perform. These include
capturing head movement using a standard camera, detecting deliberate blinks with infrared
sensors, moving the cursor based on head direction, and enabling left or right clicks through blink
gestures. The system should allow users to navigate basic computing tasks such as selecting icons,
browsing, and typing using on-screen keyboards.

Non-functional requirements address the overall quality of the system. These include reliability
(the system should perform consistently over extended use), real-time performance (minimal
latency in response to input), and accessibility (intuitive and user-friendly interface for users with
physical limitations). Additionally, the system should require minimal calibration and operate
effectively under varying lighting conditions.

The feasibility study evaluates the system from technical, operational, and financial perspectives.
Technically, it verifies the capability of the Raspberry Pi to handle image processing tasks
efficiently. Operationally, the system is designed to support individuals with mobility challenges,
offering them increased independence. Economically, the use of affordable hardware components
makes the system accessible for broader deployment.

The hardware requirements include a Raspberry Pi 3B for processing, a 720p USB camera for
capturing head movements, and infrared sensors for blink detection. Software requirements
include the Python programming language and the OpenCV library for image and video
processing. The system runs on a lightweight Linux-based operating system, ensuring stability and
resource efficiency.
In summary, this chapter lays a comprehensive foundation for developing a head and blink
movement-based system that is reliable, cost-effective, and user-centric.

3.2 General Architecture of Software

The head and blink tracker system consists of the following key components:
• Input Module: Captures real-time video frames for head movement detection and blink
recognition using the webcam.
• Processing Module: Utilizes image processing techniques to track head movements and detect

10
 deliberate blinks using infrared sensors.
• Output Module: Maps head movement to screen coordinates for cursor control and uses blinks
for click actions.
• Storage and Logging: Maintains user interaction logs for system optimization and debugging.
Flow Diagram of Gaze Tracker System

User

Camera Input

Feature Extraction

Gaze Estimation Model

Output Display

Figure 3.1 - General architecture

3.3 Requirement Specification

3.3.1 Functional Requirements

Head Detection & Tracking
• Detect and track head tilt using real-time video processing (e.g., Haar cascades or contour detection).
• Estimate head movement direction to control cursor position on screen.
• Support real-time tracking with minimal delay (latency < 50ms).
• Enable smooth and accurate cursor control based on the user's head movements.
Blink Detection
• Detect eye blinks using infrared sensors for precise tracking without environmental interference.
• Implement debounce logic to ensure only intentional blinks are registered as clicks.
• Enable left and right clicks based on blink gestures, providing a reliable method of interacting with
the system.
Calibration
• Implement a manual calibration step to ensure accurate head movement translation to cursor position.
• Allow adjustable sensitivity settings for head tilt to accommodate different users and their specific
needs.
• Use multi-point calibration for more accurate mapping of head movement to screen coordinates.
User Interaction
• Enable users to perform basic computing tasks such as icon selection, navigation, and on-screen
typing using only head and blink movements.
11
 • Allow users to interact with the system in a hands-free manner, supporting a variety of everyday
tasks like browsing and document editing.

This system is designed to enhance accessibility for individuals with physical disabilities by providing a
simple, reliable, and user-friendly interface to control a computer using only head movements and eye blinks.
3.3.2 Non -Functional Requirements

Fig -3.2- Key types of non-functional requirement

Non-functional requirements of a Gaze Tracker prediction project might include:

2.1 Performance Requirements

• Frame rate: ≥ 30 FPS for smooth head tracking and blink detection .

• Processing latency: < 50ms for real-time performance.

• Error margin: ≤ 5° of head tilt for accurate cursor control.

2.2 Usability Requirements

• User-friendly interface for easy setup and calibration.

• Minimal user training required for effective interaction.

• Support for various screen sizes and resolutions.

2.3 Security & Privacy

• No AI-based data profiling to ensure privacy.

• Local data processing with no cloud storage to protect user privacy.

• Allow users to disable tracking at any time.

12
2.4 Hardware & Software Constraints

• Requires standard webcam or infrared camera (≥ 720p resolution).

• Compatible with OpenCV, Python, for image and video processing.

• Minimal CPU and memory usage for efficient processing.

System Requirements

Functional Requirements Non-Functional Requirements

Real-time Gaze Tracking Performance & Security

Figure: 3.3 – System Requirement

3.4 System Requirements Study

The System Requirement is a critical phase in the development of the head and blink tracker
system. It involves identifying and analyzing the essential components—both hardware and
software—required to build a functional, reliable, and user-friendly hands-free computer control
interface for individuals with physical disabilities.

Purpose of the System

The main goal The main goal is to provide an assistive tool that allows physically challenged users
to interact with a computer using head movements and eye blinks, eliminating the need for traditional
input devices like a mouse or keyboard. This enhances accessibility and user independence.

Scope of the System

The system is designed to:
Detect head tilt using a USB camera.
Capture eye blinks using infrared (IR) sensors.
Translate these movements into cursor control and mouse clicks.

3.4.1 Software Requirements

1.1 Operating System

• Raspberry Pi OS (Raspbian) / Ubuntu

• Lightweight and compatible with OpenCV and Python

13
1.2 Programming Languages & Libraries

• Python 3.x – Main programming language

• OpenCV – For real-time image processing

• NumPy – For numerical operations

• Raspbian GPIO Library – For IR sensor integration

1.3 Development Environment

• Thonny / VS Code – Python IDEs for coding

3.4.2 Hardware Requirements

1.1 Processing Unit

• Raspberry Pi 3B (or later)

• Quad-core 1.2 GHz CPU

• 1 GB RAM

• HDMI output for display connectivity

• USB ports for peripherals (camera, IR sensors)

• Low power consumption for portable applications

1.2 Camera

• 720p USB Camera

• Supports 30 FPS for smooth real-time tracking

• Autofocus capability for clear eye detection

• IR-compatible (if used in low-light conditions)

1.3 Infrared (IR) Sensors

• IR LED and IR Camera Sensor (optional for low-light tracking)

• Helps in detecting pupil reflections under controlled IR lighting

• Reduces interference from ambient light

1.4 Display & Peripherals

• HDMI Monitor / LCD Screen (for real-time gaze tracking output)

14
 • USB Keyboard & Mouse (for calibration and testing)

• Power Supply (5V, 2.5A) for Raspberry Pi stability.

3.4.3 System Requirements

1.1 Environmental Constraints

• Indoor use preferred (controlled lighting conditions)

• Works with users wearing glasses/contact lenses

• IR-based tracking enhances performance in low-light

1.2 Security & Privacy

• No cloud storage (local processing for privacy)

• Data encryption for secure gaze interaction logs

• Manual control to disable tracking anytime.

3.5 System Design

3.5.1 – Introduction

The System Design for the Head and Blink Tracker System consists of the following key
components:

1. System Architecture

The head and blink tracker system is designed to enable hands-free computer control by using head
movement for cursor navigation and eye blinks for click actions. The system includes the following
main layers:

1.1 Input Layer (Data Acquisition)

• USB Camera / IR Camera: Captures real-time video frames of the user’s face and head.
• Infrared (IR) Sensors : Used for precise blink detection, especially in low-light conditions,
by identifying pupil reflections under controlled infrared lighting.

1.2 Processing Layer (Feature Extraction & Tracking)

• Preprocessing: Converts video frames into grayscale to improve contrast and highlight key
features such as the face and eyes.
• Head Detection: Uses face and eye region identification to track head movements and tilt
using techniques like Haar cascades or contour detection.
• Blink Detection: Uses infrared sensors to detect blinks by measuring the duration of eye

15
 closure.
• Head Movement Mapping: Maps the detected head tilt to screen coordinates, translating the
user’s head direction into cursor movement on the screen.
• Blink-Based Interaction: Detects intentional blinks and triggers click actions based on
predefined blink gestures (e.g., double blink for selection).

1.3 Output Layer (Interaction & Feedback)

• Cursor Movement / UI Interaction: Translates head movement into real-time cursor

movement on the screen.
• Click Actions: Triggers left and right-click events based on intentional blink gestures (e.g.,
double blink for selection).
• Real-Time Feedback: Displays cursor position and interaction status on the screen to help
the user understand their input
4 System Workflow
Step 1: Image Acquisition
• The USB camera captures video frames at a fixed frame rate to ensure smooth tracking.
• If infrared sensors are used, they work alongside the camera to enhance blink detection
accuracy, especially in low-light environments.
Step 2: Head Detection
• The system detects the face using face detection techniques (e.g., Haar cascades).
• Once the face is detected, the system identifies the eye region for tracking head movements.
Step 3: Blink Detection
• The infrared sensors are used to detect blinks by measuring the closure of the eyes.
• The system ensures that only intentional blinks are counted by implementing debounce logic,
preventing false detection.
Step 4: Gaze Mapping
• The detected head tilt is mapped to screen coordinates for cursor movement.
• The system ensures that minimal latency is maintained (less than 50ms) for real-time
interaction.
Step 5: Interaction & Control
• If the user's gaze stays focused on an area for more than a predetermined duration (e.g.,
500ms), the system triggers a click action.
• Intentional blinks are registered as left or right clicks, enabling interaction with the screen
without the need for physical input devices.
• The system provides real-time feedback on gaze and click actions, helping users know if their
commands have been registered.

16
1. System Design Diagram

User's Eye Movement

Camera Captures Frame

Image Preprocessing
Grayscale Conversion

Eye & Pupil Detection

Thresholding, Contour Det.

Gaze Mapping & UI

Cursor Control, Clicks

Output

Figure: 3.4 System Design Diagram

2. Hardware & Software Considerations

• Hardware Requirements

• Raspberry Pi 3B+ / PC for processing

• USB Camera (720p) / IR Camera

• Infrared LEDs (optional for night tracking)

• Monitor for real-time visualization

• Software Requirements

• Python 3.x

• OpenCV for image processing

• NumPy for mathematical operations

17
 CHAPTER 4
IMPLEMENTATION
This image shows a Raspberry Pi setup connected to two infrared (IR) sensors via jumper wires. The IR
sensors are likely used for detecting eye blinks or head movements, commonly implemented in hands-free
human-computer interaction systems. The Raspberry Pi serves as the processing unit, receiving signals from
the sensors and executing control logic, possibly for cursor movement or click simulation in an assistive
technology project.

Figure 4.1: Raspberry Pi setup

This image shows a wearable assistive device where IR sensors mounted on glasses are connected to a
Raspberry Pi. The setup enables hands-free computer control through eye blinks, designed for users with
physical disabilities.

Figure 4.2: wearable assistive device

18
This image shows a close-up view of a smart glasses prototype equipped with IR sensors and connected
wiring, placed in front of a monitor. The red and green indicator LEDs suggest active sensor functionality,
likely used for eye-blink or gaze-based control. The setup is part of a hands-free human-computer interaction
system powered by a Raspberry Pi in the background.

Figure 4.3: Smart glasses prototype

4.1 Code Implementation

Figure – 4.4: Code Implemented

19
Figure – 4.5: Code Implementation

Figure – 4.6: Code Implementation

20
 CHAPTER 5
TECHNOLOGY USED

The gaze tracking system is built using a combination of hardware components and software tools that
work together to achieve real-time head and eye movement detection and tracking. The hardware setup
includes a Raspberry Pi 3B, which serves as the processing unit. Its low power consumption,
affordability, and ability to handle image processing tasks efficiently make it ideal for this application.

A 720p USB camera is used for capturing the video feed, providing sufficient frame rate and resolution
for detecting head movements and facial features accurately. Infrared (IR) sensors are incorporated into
the system to assist in detecting eye blinks. These sensors enhance visibility in various lighting
conditions, ensuring reliable detection of eye movements and blinks, even in low light environments.

On the software side, the system relies heavily on Python as the programming language, leveraging its
extensive support for image processing and real-time computing. The OpenCV library plays a key role
in detecting and tracking the face and eyes. OpenCV is used for tasks such as grayscale conversion, face
detection, and the calculation of head positions.

To implement head tracking, the system uses the Haar cascade classifier for face detection. The detected
face coordinates are used to determine the center of the face, which is then mapped to the corresponding
cursor position on the screen. The cursor movement is smoothed using a simple averaging technique
and position buffering to improve accuracy.

For blink detection, the system uses IR sensors placed on the left and right sides of the user to detect
eye blinks. When a blink is detected, it triggers a corresponding mouse click (left or right), providing
hands-free interaction with the system. The system uses a debounce mechanism to prevent multiple
unintended clicks.

Unlike AI-based gaze tracking systems, this system does not rely on deep learning models, making it
lightweight and fast. Instead, it utilizes traditional image processing techniques such as Haar cascades
for face and eye detection and simple geometric transformations to map head movements to screen
coordinates. This approach ensures that the system remains efficient and does not require large datasets
or cloud-based processing.

The system is designed to be modular, allowing for future enhancements such as integrating additional
sensors, improving the gaze estimation algorithm, and potentially adding voice or gesture controls.

With the combination of affordable hardware, efficient computer vision algorithms, and real-time
processing capabilities, this gaze tracking system provides an accessible, cost-effective solution for
various applications, including assistive technologies, gaming, and smart home control.

21
 CHAPTER 6
RESULT AND DISUSSION

The Gaze Tracker system is tested with different methodology in order evaluate to its
performance on the spot cursor control and blink-based click results. The main aim was to
provide a budget-friendly, accessible solution for individuals with physical challenges, starting
them to interact with a computer using head movements.

6.1 System Accuracy and Performance Metrics

The following results were obtained during prototype testing:

Table 6.1: Testing

Metric Avg. Accuracy (%) Avg. Response Time (sec)
Head Tracking 82% 0.07
Eye Tracking 79% 0.10
Blink Detection 75% 0.05
Under optimal lighting conditions, system accuracy improved significantly:
Table 6.2: Testing Under Optimal Light Condition
Condition Head Tracking (%) Eye Tracking (%) Blink Detection (%)
Optimal Lighting 93% 90% 96%
Outdoor 88% 85% 94%

These results show that the system is highly responsive in controlled surroundings, with some
small degradation.

6.2 Discussion

The project successfully shows hands-free computer control.

Key observations include:

• Head Tracking is stable and responsive due to face landmark detection and averaging
algorithms, though occasional drift occurs during rapid or large-angle movements.
• Blink Detection using IR sensors proved effective, especially in reducing false
positives via logic. However, the positioning of the sensors on the glasses frame had to
be precise for consistent performance.
• User Experience: Feedback from test users indicated a learning curve of just a few
minutes. Most users adapted quickly to the head-movement-based cursor control,
finding it intuitive and fatigue-free.
• Limitations: Lighting variability, sensor misalignment, and differences in facial
features slightly affected accuracy and required occasional recalibration.

22
 CHAPTER 7
CONCLUSION & FUTURE WORK

7.1 CONCLUSION

The development of this low-cost, hands-free gaze-tracking system marks a significant step
forward in enhancing human-computer interaction, particularly for individuals with mobility
impairments. The system also has broader applications in areas like gaming, virtual reality, and
assistive technologies. By utilizing traditional image processing techniques, the system
efficiently detects and tracks head movements and eye blinks without the need for complex
deep learning models. This approach ensures cost efficiency, privacy, and ease of
implementation.
The system operates by detecting head movements using a Raspberry Pi 3B and a 720p USB
camera. Real-time head tracking and blink detection are achieved by analysing the camera feed
with the OpenCV library, utilizing face and eye detection techniques. The system translates
detected head movements into cursor control, while IR sensors detect eye blinks, enabling
mouse clicks. This intuitive method provides users with an alternative way to interact with a
PC, reducing the reliance on traditional input devices like keyboards and mice.
One of the key advantages of this gaze-tracking system is its affordability and accessibility.
Unlike high-end commercial eye-tracking devices, which require specialized hardware and
extensive calibration, this system offers a viable solution using readily available components,
making it more accessible to a broader audience. Furthermore, the real-time processing and
smoothing techniques applied to head position data ensure a seamless and responsive user
experience across different environments.
However, some limitations still exist. The system's accuracy can be affected by factors such as
lighting conditions, camera resolution, and head movements beyond a certain threshold. These
factors may impact tracking precision and require periodic recalibration. Addressing these
challenges will be essential in making the system more robust and adaptable for a wider range
of users and environments.

7.2 FUTURE WORK:

To further enhance the system’s capabilities and expand its potential applications, the following
improvements can be implemented in future versions:

1. AI-Based Self-Learning for Automatic Calibration:

o Implementing machine learning algorithms to enable the system to
automatically adjust calibration based on user behaviour, improving accuracy
across different users and reducing the need for manual calibration.
2. Integration of Voice Commands & Gesture Recognition:
o Combining speech recognition with gaze tracking to create a hybrid interaction
system, allowing users to perform actions using a combination of eye
movements and voice commands for enhanced interactivity.
3. Enhanced Cursor Movement & Response Time:
o Improving cursor tracking algorithms to ensure smoother, more precise
movement and responsiveness.
o Optimizing frame processing speed to reduce latency and improve the system's
23
 ability to map head movements to cursor positions in real time.
4. Expansion to Smartphones & Smart Home Devices:
o Adapting the system to support hands-free navigation on smartphones and
tablets, making it more versatile and accessible for mobile users.
o Integrating with smart home systems to allow users to control household
appliances using only eye movements and blinks.
5. Improved Environmental Adaptability:
o Enhancing the system’s ability to function effectively in varying lighting
conditions, such as outdoor settings or dim environments.
o Implementing adaptive thresholding techniques that dynamically adjust contrast
levels for better eye detection under different lighting conditions.
6. Support for Additional Accessibility Features:
o Adding additional functionalities such as text-to-speech conversion to assist
users with severe mobility limitations.
o Developing customizable user profiles to tailor interaction settings based on
individual needs, further enhancing accessibility.

24
REFERENCE

[1] Manideep, T., Upadhyay, I., Aggarwal, N., Reddy, P. G., & Das, D. (2022, April). IOT
based Eye blink detection system. In 2022 2nd International Conference on Advance Computing
and Innovative Technologies in Engineering (ICACITE) (pp. 491-494). IEEE

[2] Sreedevi, J., Reddy, M. S., & Satyanarayana, B. (2020). Eyeball Movement based Cursor
Control using Raspberry Pi and OpenCV. International Journal of Innovative Technology and
Exploring Engineering (IJITEE), 9(7).

[3] Shaik, A. R., Srinivas, C. V., & Kiran, K. (2019). Electronic Assistance for Paralysed Using
Eye Blink Detection. International Journal of Engineering Research and Applications
(IJERA), 9(08), 63-69.

[4] Shaik, A. R., Srinivas, C. V., & Kiran, K. (2019). Electronic Assistance for Paralysed Using
Eye Blink Detection. International Journal of Engineering Research and Applications
(IJERA), 9(08), 3-4.

[5] Cech, J., & Soukupova, T. (2016). Real-time eye blink detection using facial
landmarks. Cent. Mach. Perception, Dep. Cybern. Fac. Electr. Eng. Czech Tech. Univ. Prague,
1-8.

[6] Sreedevi, J., Reddy, M. S., & Satyanarayana, B. (2020). Eyeball Movement based Cursor
Control using Raspberry Pi and OpenCV. International Journal of Innovative Technology and
Exploring Engineering (IJITEE), 9(7), 1-4.

[7] Sreedevi, J., Reddy, M. S., & Satyanarayana, B. (2020). Eyeball Movement based Cursor
Control using Raspberry Pi and OpenCV. International Journal of Innovative Technology and
Exploring Engineering (IJITEE), 9(7), 2-3.

[8] Sreedevi, J., Reddy, M. S., & Satyanarayana, B. (2020). Eyeball Movement based Cursor
Control using Raspberry Pi and OpenCV. International Journal of Innovative Technology and
Exploring Engineering (IJITEE), 9(7), 1-2.

[9] Shashidhar, R., Snehith, K., Abhishek, P. K., Abhishek, B. V., & Pavitha, N. (2022, April).
Mouse Cusor Control Using Facial Movements-An HCI Application. In 2022 International
Conference on Sustainable Computing and Data Communication Systems (ICSCDS) (pp. 367-
371). IEEE.

[10]. Bernal Castano, J. R., Ota, A., Soto, E. A., Kersting, B., Dietz, A. L., Yan, Q., ... &
Domergue, G. (2025). Cursor Control using Head Tracking and System Actions using Facial
Gestures.

[11]. Stalljann, S., Wöhle, L., Schäfer, J., & Gebhard, M. (2020). Performance analysis of a
head and eye motion-based control interface for assistive robots. Sensors, 20(24), 7162.

25
[12]. Pradeep, V. (2025). Fuzzy based Computing for Disabled using 3D Head Movement.
Informatica, 49(15).

[13]. Sharma, H., Saxena, R., Kumar, S., Saini, A., Saad, M., & Faraz, M. (2022, November).
Virtual Hands: Real Time Keyboard, Desktop & Application Navigation using Gestures. In
2022 International Conference on Fourth Industrial Revolution Based Technology and
Practices (ICFIRTP) (pp. 262-267). IEEE.

[14]. Lin, C. S., Ho, C. W., Chang, K. C., Hung, S. S., Shei, H. J., & Yeh, M. S. (2006). A
novel device for head gesture measurement system in combination with eye-controlled human–
machine interface. Optics and Lasers in Engineering, 44(6), 597-614.

[15]. HeydariGorji, A., Safavi, S. M., Lee, C. T., & Chou, P. H. (2017, December). Head-
mouse: A simple cursor controller based on optical measurement of head tilt. In 2017 IEEE
Signal Processing in Medicine and Biology Symposium (SPMB) (pp. 1-5). IEEE.

26
 APPENDIX

import cv2
import numpy as np
import time
from gpiozero import DigitalInputDevice
from [Link] import Button, Controller
from picamera2 import Picamera2

# Initialize mouse controller

mouse = Controller()

# Initialize IR sensors
left_sensor = DigitalInputDevice(17) # GPIO 17 (Pin 11)
right_sensor = DigitalInputDevice(27) # GPIO 27 (Pin 13)

# Load OpenCV face detection model

face_cascade = [Link]([Link] + 'haarcascade_frontalface_default.xml')

# Screen resolution
SCREEN_WIDTH, SCREEN_HEIGHT = 1280, 720
FRAME_WIDTH, FRAME_HEIGHT = 320, 240 # Lower resolution for better performance

# Cursor movement improvements

prev_x, prev_y = SCREEN_WIDTH // 2, SCREEN_HEIGHT // 2
smoothing_factor = 0.9 # Increased smoothing
sensitivity = 0.25 # Reduced sensitivity
position_buffer = [] # Increased buffer size for averaging
BUFFER_SIZE = 5 # Number of positions to average
27
# Blink detection variables
debounce_delay = 0.2
left_prev = left_sensor.value
right_prev = right_sensor.value

# Initialize PiCamera2
picam2 = Picamera2()
config = picam2.create_preview_configuration(main={"size": (FRAME_WIDTH, FRAME_HEIGHT)})
[Link](config)
[Link]()

def map_head_to_screen(head_x, head_y, frame_width, frame_height):

x_span = frame_width * sensitivity
y_span = frame_height * sensitivity

screen_x = [Link](frame_width - head_x,

[frame_width/2 - x_span/2, frame_width/2 + x_span/2],
[0, SCREEN_WIDTH])

screen_y = [Link](head_y,
[frame_height/2 - y_span/2, frame_height/2 + y_span/2],
[0, SCREEN_HEIGHT])

return int(screen_x), int(screen_y)

try:
while True:
frame = picam2.capture_array()
28
gray = [Link](frame, cv2.COLOR_BGR2GRAY)

# Stabilized face detection parameters

faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=8, minSize=(120, 120))

if len(faces) > 0:
x, y, w, h = faces[0]
head_x, head_y = x + w//2, y + h//2

position_buffer.append((head_x, head_y))
if len(position_buffer) > BUFFER_SIZE:
position_buffer.pop(0)

avg_x = sum(p[0] for p in position_buffer) / len(position_buffer)

avg_y = sum(p[1] for p in position_buffer) / len(position_buffer)

mouse_x, mouse_y = map_head_to_screen(avg_x, avg_y, FRAME_WIDTH, FRAME_HEIGHT)

# Enhanced smoothing
mouse_x = int(prev_x * (1 - smoothing_factor) + mouse_x * smoothing_factor)
mouse_y = int(prev_y * (1 - smoothing_factor) + mouse_y * smoothing_factor)

[Link] = (mouse_x, mouse_y)

prev_x, prev_y = mouse_x, mouse_y

# Improved blink detection with feedback

left_current = left_sensor.value
if left_prev and not left_current:
print("Left blink detected - Left sensor activated")
[Link]([Link], 1)
29
[Link](debounce_delay)
left_prev = left_current

right_current = right_sensor.value
if right_prev and not right_current:
print("Right blink detected - Right sensor activated")
[Link]([Link], 1)
[Link](debounce_delay)
right_prev = right_current

[Link]("Optimized Head Control", frame)

if [Link](1) & 0xFF == ord('q'):
break

finally:
[Link]()
[Link]()

30
 RESEARCH PAPER 1
GAZE TRACKER
“Controlling cursor using head”

Kunal Yadav, Manav Adhana, Satendra Kumar, Maneesha, Khushi

CSE, Meerut Institute of Engineering and Technology
[Link].2021@[Link], [Link].2021@[Link],
[Link].2021@[Link], [Link].2021@[Link],
satendra04cs41@[Link],

Abstract –
Many people with disabilities face barriers in
This research paper describes the traditional ways of using a computer system. The
implementation of raspberry pi technology to limitations in movement experienced by some
build persons with disabilities eye movement individuals can limit or eliminate the use of point
controlled pc system. With head and eye blink devices such as the mouse or the keyboard. While
movements, the system enables the user to assistive innovations have advanced throughout
control the computer cursor providing physically the long term, many existing arrangements will
challenged persons with a hands-free interface. quite often zero in on solitary functionalities, for
The architecture features a Raspberry Pi 3B, a example, voice orders or versatile switches. Sadly,
720p USB Camera and a number of infrared these instruments seldom offer a complete, sans
sensors – for the purpose of blink detection. The hands interface that incorporates numerous info
camera registers head tilting, which enables the strategies, prompting a divided client experience.
AT to change the position of the pointer, while This highlights the need for a durable, instinctive
the IR sensors respond to blinks and simulate framework that permits clients to control their PCs
clicking of the mouse. The methodology consistently utilizing head developments and eye
comprises of hardware interfacing with software squints [1][2].
programming in Python and OpenCV and
followed by designing of experiments for Ongoing headways in innovation, especially in the
calibrations and testing. fields of PC vision and AI, have opened additional
opportunities for assistive gadgets. Eye-global
The evidences suggest that the system enables the positioning frameworks have shown guarantee in
user to differentiate between intention and reflex empowering sans hands association, permitting
blinks and as a consequence the user is able to clients to explore computerized conditions with
control the cursor accurately. The consequences their look [3][4]. Nonetheless, a considerable lot
of this development are the technologies that of these frameworks require costly equipment or
focus on accessibility instruments improvement complex arrangements, which can restrict their
providing a user with increased degree of openness. Moreover, existing arrangements
independence despite physical limitations. In the frequently don't successfully separate among
future additional features could include machine purposeful and reflexive developments, which can
learning algorithms for self-calibration and prompt disappointment for clients attempting to
additional capabilities for meaningfully enriched accomplish exact control [5].
physiogenic gesture typing.
To overcome this issue, our examination presents
Keywords - : Eye Tracking, Raspberry Pi, an eye-controlled PC framework that uses
Assistive Technology, Cursor Control, Blink Raspberry Pi innovation to make a reasonable and
Detection easy to understand interface. This framework
consolidates head following and flicker
identification to reenact mouse developments and
1. Introduction snaps, giving a sans hands insight to clients with
31
actual incapacities [6]. By utilizing minimal
expense parts and open-source programming, the This divided methodology highlights a basic
framework means to be broadly open, permitting examination hole: the requirement for brought
people to communicate with their PCs in a way together frameworks that consistently join eye
that suits their one of a kind necessities. following, head following, and squint discovery.
Such mix can convey ongoing criticism and
natural control, engaging clients with actual
This paper describes the development of an eye- incapacities to connect with PCs all the more
controlled PC, including its hardware, software successfully [5][8].
development, and user evaluation processes. A
general context data flow diagram is presented in Our proposed framework tends to this hole by
order to detail the core processes of the system, coordinating these components into a minimal
such as input data acquisition through sensors and expense, easy to use stage. Using Raspberry Pi, IR
camera, its processing related to the cursor sensors, and open-source programming, it intends
movement and simulated click. The results to advance inclusivity and freedom by offering a
presented in this paper state that the system can strong, assistive arrangement [1][4].
track head motions and recognizes intentional
blinks accurately, which provides comfortable and 3. Research Objectives
quick interactions with the user. Insomuch as
addressing the problems of people with limited To develop an affordable and accessible
mobility, the research may be regarded as an assistive technology system that enables
additional attempt in the direction of improving individuals with physical disabilities to control a
the inclusion and the level of autonomy in the computer cursor using head movements and eye
information world. blinks.
To integrate head tracking, eye tracking, and
blink detection into a unified system for seamless
2. Literature Review interaction.
To improve the accuracy and reliability of
Assistive technologies are progressing especially cursor control by differentiating between
in the fields of computer vision and machine intentional and reflexive blinks using infrared (IR)
learning which have been key in changing the sensors.
landscape for comfortable use by individuals with To evaluate the system's performance in
a physical disability. Eye-tracking systems used to different environmental conditions (e.g., optimal
be built around expensive, specialized equipment, lighting vs. outdoor settings) and measure its
today machine learning can detect eye movements accuracy and response time.
using typical supported webcams. The To enhance user experience by reducing false
consequence of this evolution has been the positives in blink detection and ensuring smooth
increase in the ease of adoption of eye-controlled cursor movements.
interfaces [2][3]. To provide a cost-effective and open-source
solution using Raspberry Pi, OpenCV, and IR
Past eye following, coordinating head following sensors for easy implementation and scalability.
has arisen as a correlative methodology, To explore future enhancements such as
improving accuracy in cursor control. Double machine learning for adaptive calibration and
modular frameworks consolidating head and eye additional gesture-based controls.
following give better precision and a more natural
client experience [6][7]. In any case, existing 4. Body: (Proposed methodology/ Design and
frameworks frequently need reconciliation with implementation of proposed system)
extra information techniques, for example, squint
discovery for recreating mouse clicks, coming This examination presents a sans hands
about in independent arrangements that limit ease framework intended to assist people with actual
of use for clients with shifting requirements [4][7]. incapacities connect with PCs utilizing head
32
developments and eye squints. The framework different lighting conditions, making it
uses a standard webcam and infrared (IR) sensors appropriate for various conditions.
mounted on a custom 3D-printed goggle edge to
follow head developments and recognize b.2 Blink Detection Module: Dissimilar to
purposeful flickers. These data sources are then conventional frameworks that depend exclusively
converted into mouse cursor developments and on video handling, our venture integrated infrared
snaps, making an open and easy to understand sensors mounted on a 3D-printed goggle edge to
interface. The accompanying areas portray the distinguish squints. These infrared sensors assume
framework's overall design, modules, a basic part in distinctive deliberate squints from
calculations, and the use of these strategies to compulsory eye developments [3][4]. By
construct our task. checking the mirrored IR light from the eyes, the
framework can precisely identify the snapshot of
a) General Architecture a flicker. The mix of IR sensors and picture
The framework engineering is measured and handling calculations permits us to limit
contains numerous parts that cycle constant misleading up-sides and guarantee that main
information. The initial step includes catching purposeful flickers are recognized as mouse
video from a standard webcam, trailed by the clicks. This double methodology altogether
investigation of head developments and flicker upgrades the client experience, taking into
location. These activities are converted into cursor consideration exact and dependable command
developments and snaps, giving clients an over the framework.
instinctive method for controlling their PC. The
framework's plan guarantees it is open on any c) Algorithms
gadget with a webcam and a web association,
killing the requirement for particular equipment or c.1 Head Tracking Algorithm: The head
programming establishments. Clients connect following calculation in our venture processes
with the framework utilizing basic head signals video outlines caught by the webcam to recognize
and deliberate squints, making the framework facial milestones and compute the direction of the
simple to utilize and exceptionally successful for client's head. Utilizing dlib's facial milestone
individuals with versatility challenges. These identification model, we distinguish the general
strategies were applied in our task by coordinating places of facial elements (e.g., eyes, nose, and
them into a product arrangement that processes mouth), which are then planned into cursor
video edges and sensor information progressively, developments. The calculation ceaselessly
guaranteeing smooth and responsive control. refreshes the cursor's position in view of head
developments, giving constant control. This
b) Module Descriptions technique guarantees smooth and exact
connections, making it reasonable for clients who
b.1 Head Tracking Module: The head following depend on head developments to control the PC.
module in our framework uses OpenCV and dlib,
which are deep rooted libraries for facial c.2 Blink Detection Algorithm: The squint
milestone location and PC vision applications identification calculation in our venture depends
[1][2]. The webcam catches video outlines, which on infrared (IR) sensors, which are mounted on a
are handled to recognize key facial milestones like custom 3D-printed goggle outline. These IR
the eyes, nose, and mouth. The framework works sensors identify changes nearby the client's eyes
out the general places of these milestones to when a squint happens, without utilizing picture
decide the client's head direction and make an handling methods. The framework enrolls these
interpretation of it into cursor development. This fast changes as flickers in light of explicit edges,
strategy takes into account ongoing changes, guaranteeing that main purposeful squints are
empowering clients to definitively explore their distinguished and considered clicks. This
screens. approach limits bogus up-sides from compulsory
To guarantee dependable execution, the head squints and improves the framework's precision
following module was enhanced to work in and responsiveness. By utilizing IR sensor
33
innovation, the framework offers a more direct Avg. Accuracy Avg. Response
and dependable strategy for distinguishing eye Metric (%) Time (sec)
squints contrasted with customary picture based
frameworks, which can be delicate to lighting and
other ecological elements. The mix of head Head Tracking 82% 0.07
following and squint recognition furnishes clients
with a consistent encounter, permitting them to Eye Tracking 79% 0.10
control the cursor and perform clicks with head
developments and flickers alone. The workflow of Blink
this design will be similar to the fig (1.1). Detection 75% 0.05
This procedure fundamentally lessens the
intricacy and cost of the framework while keeping
an elevated degree of accuracy and convenience (Table 1) Performance metrics: Accuracy and
for people with actual inabilities. response time for head tracking, eye tracking, and
blink detection systems.
Flowchart

Head Eye Blink

Tracking Tracking Detection
Accuracy Accuracy Accuracy
Condition (%) (%) (%)

Optimal
Lighting 93% 90% 96%

Outdoor 88% 85% 94%

(Table 2) Performance metrics: Accuracy and

response time for optimal lighting and outdoor
systems

Figure 1.1; A visual representation of the order

process, starting with the activation of an IP
camera and IR sensor, detecting head movements
and blinks to simulate cursor actions and clicks for
efficient control.
.
4. Result analysis/ result discussion

34
 clients communicated fulfilment with the head
following and flicker identification modules, a
couple of mentioned improvements in cursor
responsiveness and extra elements, for example,
voice order reconciliation, to make the framework
more flexible.

5. Conclusion
In rundown, this examination tends to the test of
fostering a coordinated, continuous eye-controlled
PC framework that improves connection for
people with actual handicaps. The essential main
thing in need of attention is the absence of a strong
arrangement that joins head following, eye
following, and squint location, which together
work on the openness and convenience of PC
interfaces.
The trial discoveries exhibit that this framework
accomplishes high exactness, quick reaction
times, and instinctive client association across
different natural circumstances. By using
Prototype setup: Raspberry Pi with connected IR progressed calculations for head following, eye
sensors and webcam for head and blink detection following, and squint location, the framework
system. offers a bound together and viable arrangement
that works with continuous control of
computerized conditions. This joining spans
The after effects of the exploratory tests were existing holes in assistive innovation, enabling
incorporated to assess the presentation of every clients to explore their processing assignments no
module. Head following shows the high rate in sweat and certainty.
various conditions, with possible negligible By and large, the framework's capacity to
misfortune in carrying out regular utilization precisely decipher and answer client
situations. The framework execute dependably in developments progressively highlights its true
both typical and low-light situations, furnishing an capacity as an extraordinary instrument, adding to
excellent client experience even with a standard a more comprehensive and open mechanical scene
webcam. for people confronting actual difficulties.
Flicker identification through IR sensors was Albeit the framework shows extensive
assessed in various preliminaries, incorporating commitment, future enhancements could include
members with changing facial designs and squint upgrading the scope of the IR sensors to oblige
frequencies. The 3D-printed goggle outline clients with changing head positions. The ongoing
guaranteed appropriate situating of the sensors, framework is intended to work inside a restricted
adding to high identification rates and limiting head development reach, and future variants could
bogus up-sides. The IR-based identification profit from consolidating a more extensive scope
strategy ended up being quicker and more solid of movement for further developed adaptability.
than picture handling based frameworks, as it
wiped out normal difficulties like facial Further progressions could incorporate the
impediments or lighting varieties. coordination of voice orders or motion controls to
supplement the current info strategies. These
Client input was additionally gathered after the augmentations could altogether upgrade the
tests, and 75% of members detailed that the framework's flexibility, considering more refined
framework was natural to utilize. While most associations, especially for clients with restricted
35
head or eye portability. Also, integrating versatile [7] Sreedevi, J., Reddy, M. S., & Satyanarayana,
cursor speed and responsiveness could B. (2020). Eyeball Movement based Cursor
additionally further develop the client experience Control using Raspberry Pi and
by fitting the point of interaction to individual OpenCV. International Journal of Innovative
necessities. Technology and Exploring Engineering
(IJITEE), 9(7), 2-3.
Future work will likewise zero in on refining the
framework's alignment cycle to guarantee it is fast [8] Sreedevi, J., Reddy, M. S., & Satyanarayana,
and simple for clients to set up without the B. (2020). Eyeball Movement based Cursor
requirement for particular specialized Control using Raspberry Pi and
information. OpenCV. International Journal of Innovative
Technology and Exploring Engineering
References (IJITEE), 9(7), 1-2.
[1] Manideep, T., Upadhyay, I., Aggarwal, N.,
Reddy, P. G., & Das, D. (2022, April). IOT
based Eye blink detection system. In 2022 2nd
International Conference on Advance
Computing and Innovative Technologies in
Engineering (ICACITE) (pp. 491-494). IEEE

[2] Sreedevi, J., Reddy, M. S., & Satyanarayana,

B. (2020). Eyeball Movement based Cursor
Control using Raspberry Pi and
OpenCV. International Journal of Innovative
Technology and Exploring Engineering
(IJITEE), 9(7).

[3] Shaik, A. R., Srinivas, C. V., & Kiran, K.

(2019). Electronic Assistance for Paralysed
Using Eye Blink Detection. International
Journal of Engineering Research and
Applications (IJERA), 9(08), 63-69.

[4] Shaik, A. R., Srinivas, C. V., & Kiran, K.

(2019). Electronic Assistance for Paralysed
Using Eye Blink Detection. International
Journal of Engineering Research and
Applications (IJERA), 9(08), 3-4.

[5] Cech, J., & Soukupova, T. (2016). Real-time

eye blink detection using facial landmarks. Cent.
Mach. Perception, Dep. Cybern. Fac. Electr.
Eng. Czech Tech. Univ. Prague, 1-8.

[6] Sreedevi, J., Reddy, M. S., & Satyanarayana,

B. (2020). Eyeball Movement based Cursor
Control using Raspberry Pi and
OpenCV. International Journal of Innovative
Technology and Exploring Engineering
(IJITEE), 9(7), 1-4.

36
 RESEARCH PAPER 2

Eye and Head Movement-based Cursor Control System: A Survey

Kunal Yadav, Manav Adhana, Satendra Kumar, Maneesha, Khushi

Departement of CSE, Meerut Institute of Engineering and Technology, Meerut, U.P, India

[Link].2021@[Link], [Link].2021@[Link], satendra04cs41@[Link],

[Link].2021@[Link], [Link].2021@[Link],

physio genic gesture recognition are examined

for their potential to enhance system autonomy
Abstract- Accessibility for people with physical
and usability.
disabilities has really gotten better because of new
technology. This review looks at a control system
that uses head movement to operate computers. Enhanced techniques and advanced sensor
The development of eye and head movement- fusion methods could significantly enhance user
based cursor control systems is the focus of this experience and reliability. However, there are still
review, We focus on hardware setups, computer challenges regarding user-specific customization,
vision techniques, & ways to interact. Max project environmental adaptability, and processing
utilizes Raspberry Pi, infrared sensors, & USB speed. Our review paper shows the ability to
cameras in order to track head movements. understand the starting state of research,
Python includes OpenCV which allows real-time highlights key limitations, and discusses possible
processing , that makes the system quicker & future developments for
more refine henceforth some challenges like
need to fix settings and figuring out the difference
hands-free computing interfaces that are more
between intentional reflexive ones. Various
user-friendly and dependable.
implementation include Raspberry Pi, IR sensor,
and camera to track head motions at instant made
possible by the use of Python and OpenCV, shows Keywords—: Eye Tracking, Raspberry Pi,
system results that improve exactness. Assistive Technology, Cursor Control, Blink
Detection
The need for adaptive interconnectivity and the
difficulty of achieving consistent nationwide I. Introduction
performance using adjustable blinks are
highlighted by a critical analysis of existing
solutions. There are traditional computer systems, many
people with physical challenges face relative
difficulties. Access barriers, which minimize
Expanding methodologies in gaze tracking and online interaction, can be created by people who

37
are unable to use basic input devices like a For instance, Álvarez et al. (2013) presented an
keyboard or mouse. gaze tracking system for human-computer
Whereas, different technologies have been interaction using low-cost webcams,
standardized over the years, many existing demonstrating promising accuracy for basic
solutions focus on singular functionalities, such as pointer control. In a similar vein, Villanueva and
voice commands or adaptive switches. Henceforth Cabeza (2007) emphasized vision-based systems
, these techniques often fail to provide the ability to for eye tracking that could be utilized with
understand, hands-free interface that smoothly commercial-grade cameras in real-time
integrates multiple input methods, resulting in a environments. More recent innovations, such as the
disjointed user experience. This demonstrates the Eye Writer project (Golan et al., 2010), made use
need for a more robust and user-friendly system of open-source software and hardware to enable
that allows users to effectively control their paralyzed people to communicate and draw with
computers through head movements. The eye movements. Maazaoui et al. also looked into
percentage for these devices have increased as a mouse click simulation systems that use blink
result of improved techniques over the time period, detection. (2018), who integrated infrared sensors
especially in the fields of computer vision. Our with a neural network classifier to distinguish
head tracking systems have shown promise in intentional blinks. These related studies have laid
making it possible for users to use their gaze to the foundation for accessible, affordable, and real-
navigate online environments and allow for free time assistive technologies, aligning well with the
interaction. proposed Raspberry Pi-based system using Python
Moreover , many of these systems depend on and OpenCV. The current work builds on these
expensive hardware or difficult useable processes advancements by incorporating head tilt and blink-
which makes them less accessible to a large based controls using minimal hardware, which
number of people. Additionally, it is still difficult contributes to the growing field of hands-free
to tell between movements that are planned and interfaces for physically challenged users.
those that are automatic, which frequently results
in inaccurate cursor control and user frustration.
II. Related Work
Eye and head movement-based cursor control
systems are the subject of this review, which
focuses on hardware configurations, software Several prior research efforts have explored
methodologies, and user experience assistive technologies leveraging eye and head
considerations. In order to highlight the benefits, movements for computer interaction, particularly
drawbacks, and practical applicability of various targeting individuals with physical disabilities. For
assistive solutions, a comparative analysis is instance, Álvarez et al. (2013) presented an eye-
carried out. In addition, new developments in gaze tracking system for human-computer
accessibility and user autonomy are examined in interaction using low-cost webcams,
the paper, such as the application of machine demonstrating promising accuracy for basic
learning to adaptive calibration and gesture-based pointer control. Similarly, Villanueva and Cabeza
typing. This review aims to contribute to the (2007) emphasized vision-based systems for eye
ongoing development of more effective hands-free tracking that could be implemented in real-time
computing interfaces for people with physical environments using commercial-grade cameras.
disabilities by identifying research gaps and More recent developments such as the Eye Writer
evaluating future directions. project (Golan et al., 2010) employed open-source
hardware and software to enable paralyzed
individuals to draw and communicate using eye
II. CONNECTED WORK
movements. Furthermore, systems utilizing blink
detection for mouse click simulation were studied
Several previous studies have looked into assistive by Maazaoui et al. (2018), who integrated infrared
technologies that use eye and head movements to sensors with a neural network classifier to
interact with computers, especially for people with distinguish intentional blinks. These related studies
physical disabilities. have laid the foundation for accessible, affordable,

38
and real-time assistive technologies, aligning well 2. 720p USB Camera
with the proposed Raspberry Pi-based system
using Python and OpenCV. The current work
Captures real-time video of the user's face and
builds on these advancements by incorporating
head movements. The camera feeds are used for
head tilt and blink-based controls using minimal
head tracking to move the cursor based on the tilt
hardware, which contributes to the growing field of
and orientation of the user’s head.
hands-free interfaces for physically challenged
users.
3. Infrared (IR) Sensors
III. PROBLEM STATEMENT
Detect eye blinks using infrared light. These
sensors are crucial for distinguishing intentional
There are many people who often face hardship and
blinks (used for mouse clicks) from natural or
challenges while having a interaction with
reflexive blinks.
traditional computer input and output devices like
mouse and keyboard. These basic peripherals
require fine motor control and hand coordination, 4. Python
which may not be possible for individuals with
conditions such as paralysis, muscular dystrophy,
or spinal cord injuries. Even though many different The primary programming language used to
different technologies that exist to address these develop the control logic, interface with hardware
limitations out of which of them are extra- components, and manage real-time data
ordinarily expensive, require basic hardware, or are processing.
not readily available in low-resource settings. This
creates a digital divide, limiting opportunities for 5. OpenCV (Open Source Computer Vision
education, employment, and communication for Library)
those affected. There is a need for a cost-
effectiveness, more accessible, and user-friendly
solution is really very important. Addressing this A powerful computer vision library used for
gap, the proposed system introduces a low-cost, processing video input, detecting facial features,
hands-free computer interaction method utilizing tracking head movements, and integrating blink
head movements for cursor navigation and eye detection for cursor control.
blinks for mouse click simulation. This approach
uses affordable components such as a Raspberry
Pi, USB camera, and infrared sensors, offering an
innovative alternative to high-end systems while
promoting digital accessibility and independence
for people with physical impairments.

.
IV. METHODOLOGY
A. TECHNOLGIES USED
This study presents a comprehensive survey of
1. Raspberry Pi 3B gaze and head-movement-based cursor control
systems in Human-Computer Interaction (HCI). It
aims to systematically analyze publication trends,
A compact, low-cost, and energy-efficient technological techniques, and tools used in the
single-board computer used to process input data development of gaze tracker systems. The research
from sensors and run the control software. It serves covers work published from *2006 to 2025*, with
as the central controller for the system. a primary focus on methods utilizing head and eye

39
movements to control computer input devices.

- IEEE (4 papers)

- MDPI (1 papers)

- Elsevier (1 paper)

- IRJET (2 papers)

- Others (12 papers)

Number of Retrieved Papers Year
wise
This data was compiled from both journal and
15 conference papers, as visualized in Figure 1 and
10 Figure 2. The highest number of papers were
published in the 2021–2025 period, indicating a
5 growing interest in gaze-based systems due to
0 advancements in machine learning and assistive
2006-2010 2011-2015 2016-2020 2021-2025 technologies.
Journal Conference
1. Data searching techniques

Figure 1 Number of Retrieved Papers Year wise To identify and collect relevant literature for this
review, we applied a structured keyword search
strategy based on a combination of primary terms,
synonyms, related expressions, and alternative
spellings. This method ensured a comprehensive
exploration of studies pertinent to our topic on
Category Wise Published Papers
cursor control systems using eye and head
Symposium
movements. Keywords were selected to reflect
both the input modalities (e.g., eye movement,
Workshop
head tracking) and the target system functionalities
Journal (e.g., cursor control, mouse pointer movement).
Conference Boolean operators were used to construct effective
0 5 10 15 20 search queries: the operator ‘OR’ was used to link
synonyms and variations of a concept, ensuring
Series2 Series1 broad coverage, while the operator ‘AND’ was
used to connect distinct conceptual groups. As a
result, we created a set of Boolean search strings,
Figure 2 Category wise Published Papers summarized in Table 1, which included
combinations such as "Eye and Head Movement"
AND "Cursor Control System", and "Eye Gaze
A total of 22 research papers were included based
Tracking" OR "Control Cursor Movement". Some
on their relevance to gaze or head-controlled cursor
articles demonstrated the effectiveness of our
interfaces. Sources include well-known publishers
keyword strategy in capturing diverse research
such as:
efforts on non-traditional human-computer
interaction modalities. For this strategy, we created

40
the search string that is shown in Table 1.

Table 1 Possible search strings

TSR search strings Boolean operation SPL search strings

Eye and Head

Movement And Cursor Control System

Eye-Ball Movement Cursor Control System

EYE GAZE CONTROL CURSOR

TRACKING MOVEMENT

Mouse Pointer
Control Using Eye Movements

2. Data gathering technique Table 2 Data gathering from different digital

sources
To maintain a wider coverage in the area of
HCI, we restricted our search in our study SNO. Publisher name
work to 2006 through 2025. In order to find
the content using the keywords provided
above, we looked through a variety of 1 Technoscience Academy
international conferences, journals and
workshops. We also looked through related 2 clareus Scientific
white papers, technical (industrial) reports,
books, and theses that contained
information about HCI. An earlier stage in 3 IEEE
this field is searching with the keyword. We
created a search string before gathering 4 IJARCS
information pertinent to our task. A search
option is offered by the literary digital
resources as shown in Table 2. 5 IJSRR

6 INTJECSE

7 IRJET

8 MDPI

41
 to space limitations, this article doesn’t involve in
data extraction. Three researchers have
9 MZUJHSS
participated in the verification and data extraction.
All the papers, which were selected, have been
10 NN qualitatively studied. In order to synthesise the
data, the meta-analysis method is inappropriate.
We have manually reviewed them and merged the
11 Others extracted data. We found the problem from each
and every paper and record it in the notebook.
3. Selection of study data Table 3 lists the papers used to complete the survey
in the next few years.

Search terms were entered into the search engines

that were displayed in the prior part to discover Table 3 Number of published papers in each year
similar information. The process of screening is 2006–2025.
used to choose the studies, and it includes filters
like finding pertinent studies by using the Year of Publication Number of Paper Published
aforementioned keywords, correcting the search
results by removing study titles, and then excluding
the studies after carefully reading their abstract and 2006-2010 1
conclusion sections. Include studies if they meet
the following requirements:
2011-2015 2

• Focus on gaze-based cursor control or

pointer interaction 2016-2020 6

2021-2025 13
• Use of eye or head movement for input

• Implementation with a defined

technique or tool (e.g., OpenCV, IR
Sensor, EEG) V. Survey results and discussion

• Clearly stated methodology and Over the past decade, significant research has been
application domain conducted to explore hands-free Human-Computer
Interaction (HCI) techniques using head
movements, facial gestures, eye tracking, and
Table 4 shows that highest number of research
related sensor-based approaches. Based on the
papers was published in conference and journals
gathered data (see publication and technique
but symposium and workshop has no papers those
tables), it is evident that researchers have
are related to HCI.
consistently explored various algorithmic and
hardware approaches to improve cursor control
All 22 papers have been fully researched to extract systems, especially for users with physical
relevant data. To extract and store data, we use a impairments.
predefined form that contains many attributes. Due
and 2025, reflecting a recent surge in interest.
Early work (pre-2010) is minimal, indicating the
As shown in the table below, the highest number
novelty of this HCI subdomain.
of journal publications occurred between 2021

42
Table 4 TSR-based work published in different years (2006–2025)

2006- 2011- 2016-

Venue Year 2010 2015 2020 2021-2025 Total

Conference 0 0 2 2 4

Journal 1 2 4 11 18

Workshop 0 0 0 0 0

Symposium 0 0 0 0 0

Total 1 2 6 13 22

Figure 3 Publication-based selection of articles

Category Wise Published Papers
Symposium Research papers for various conferences vary. As
Workshop indicated in Table 5, these conferences are quite
Journal
well attended, but the IEEE is the most well-known
conference that publishes research papers on
Conference
human computer interface. While Table 5 also lists
0 5 10 15 20 the names of other publications, one of them is the
well-regarded Journal of Advance Research in
Series2 Series1
Computer science and Engineering.

43
Table 5 Publication distribution

Publication 2022 International Conference

Type on count
Sustainable Comp
Communication Systems (ICSCDS)
2017 IEEE Signal Processing in Medicine and Biology Symposium (SPMB) Conference 1
International Journal of Advanced Research in Computer S
ScienceDirect Journal 1
International Journal of Early Childhood Special Education
Fourth Industrial Revolution Based Technology and Practices (ICFIRTP),
International Conference on International Research Journal of Engineering
Conference 1 and Technol

Clareus Scientific ScienceJournal

International Research Journal of Engineering and Technology and Engineering
1

Informatica International Journal of Scientific

Journal and Research
1 Publication

Sensors International Journal of Journal

Scientific Research
2 in Science,
Technology
International Journal for Research in Engineering Application & Management Journal 1
Journal of Advance Research in Computer science and Eng
Technical Disclosure Commons Journal 1
IEEE TRANSACTIONS ON NEURAL SY
REHABILITATION ENGINEERING
UACEE International Journal of Computer Science and its Applications Journal 1

Mizoram University Journal of Humanities & Social Scienc

Ijraset Journal For Research in Applied Science and Engineering Technology Journal 1

Proceedings of TENCON 2018 - 2018 IEEE Region 10 Con

The key studies listed below were used to finish our Control Interface for Assistive Robots’, was
survey on HCI. Table 6 includes the following published in the MDPI journal. The other, titled
three fields: reference, title, and venue, together ‘Virtual Hands: Real Time Keyboard, Desktop &
with the year of publication and page numbers. One Application Navigation using Gestures’, was
of these papers, titled ‘Performance Analysis of a published in the IEEE. Other than these two
Head and Eye Motion-Based publications are based on experimental research
with distinct recommended approaches.

44
Table 6 Primary studies selection

References title Venue

Khushal Sancheti et al. Hands-free Cursor Control using Intuitive TENCON'18

Head Movements and Cheek Muscle Twitches

Vadde Veeresh Human Computer Interaction based Head MZUJHSS'24

Controlled Mouse

Wahyu Rahmaniar Touchless Head-Control (THC): Head Gesture IEEE TNSRE'22

Recognition for Cursor and Orientation
Control

Hemraj Ashtankar MOUSE CONTROL USING HEAD NN Publisher'15

MOVEMENT

Akshai P Sajeev Cursor Control Using Head and Hand Technoscience

Movement Academy'16

Vaibhav Nangare Controlling Mouse Motions Using Eye Blinks, IJSRR'16

Head Movements and Voice Recognition

G Tanusha Enabling Cursor Control through Eye clareus Scientific'25

Movement using Hidden Markov Model

.Madan Venkat Sai Facial-Expression Based Mouse Cursor IRJET'23

Control for Physically Challenged Individuals:
A Human Computer Interaction (HCI)
Approach.

Alberto [Link]- Controlling a Mouse Pointer with a Single- MDPI Sensors'21

Cantero Channel EEG Sensor

Mrs. Srilatha Puli HUMAN COMPUTER INTERACTION INTJECSE'23

BASED HEAD CONTROLLED MOUSE

Aniket Kudale Mouse Navigation Control Using Head IJARCS'12

Motion

Sashidhar R Mouse Cusor Control Using Facial IEEE ICSCDS'22

Movements - An HCI Application

45
 Saurabh Barve A Portable Wireless Head Movement IJRASET'23
Controlled Human-Computer Interface for
People with Disabilities

Yashaswita bhoir Hands-free Cursor Control UACEE'24

Jonathan Ray Bernal Cursor Control using Head Tracking and Defensive Publications
Castano System Actions using Facial Gestures series'25

Anurag Singh COMPARATIVE STUDY OF HAAR AND IJREAM'17

ADABOOST ALGORITHMS FOR
CURSOR CONTROL USING HEAD AND
EYE MOVEMENT FOR DISABLES

Sarah Stalljann Performance Analysis of a Head and Eye MDPI Sensors'20

Motion-Based Control Interface for Assistive
Robots

Pradeep V Fuzzy based Computing for Disabled using 3D Informatica'23

Head Movement

G. Abhishek Mouse Cursor Control Hands Free Using Deep IRJET'23

Learning

Harshit Sharma Virtual Hands: Real Time Keyboard, Desktop IEEE ICFIRTP'22
& Application Navigation using Gestures

Chern-Sheng Lin A novel device for head gesture measurement Elsevier OLE'06
system in combination with eye-controlled
human–machine interface

Ali heydarigorji HEAD-MOUSE: A SIMPLE CURSOR IEEE SPMB'17

CONTROLLER
BASED ON OPTICAL MEASUREMENT
OF HEAD TILT

researchers to explore a range of algorithms and

hardware—from simple webcams to EEG-based
Controlling the cursor through non-traditional
systems.
input such as head movements, facial gestures, or
brainwave patterns has emerged as a key area of
interest in Human-Computer Interaction (HCI), The exponential variation in sensor capabilities,
especially for assistive technology. The growing computing
demand for hands-free interaction has led

46
environments, and user-specific needs makes contributions in this area, highlighting the
robust cursor control design complex. To technique used, the problem addressed, and the
address these challenges, researchers have proposed solution.
proposed different models using computer vision,
signal processing, machine learning, and wearable
devices. The table below summarizes major

Table 7 Test suite reduction/test case reduction-based research work details.

Authors Paper title Technique used Tools used Domain

Hands-free Cursor Control using

Khushal Intuitive Head Movements and Cheek Assistive hands-free
Sancheti et al. Muscle Twitches Technology HCI device HCI

Human Computer Interaction based

Vadde Veeresh Head Controlled Mouse Python webcam HCI

Touchless Head-Control (THC):

Wahyu Head Gesture Recognition for Cursor Assistive RGB
Rahmaniar and Orientation Control technology camera HCI

Hemraj MOUSE CONTROL USING HEAD

Ashtankar MOVEMENT haar algorithm C# HCI

Cursor Control Using Head and Hand

Akshai P Sajeev Movement OpenCV webcam HCI

Controlling Mouse Motions Using

Vaibhav Eye Blinks, Head Movements and
Nangare Voice Recognition Haar cascade webcam HCI

Enabling Cursor Control through Eye

Movement using Hidden Markov
G Tanusha Model Markov Model webcam HCI

Facial-Expression Based Mouse

Cursor Control for Physically
Challenged Individuals: A Human
Madan Venkat Computer Interaction (HCI) Raspberry Pi and
Sai Approach. OpenCV webcam HCI

Alberto electro-
Controlling a Mouse Pointer with a
[Link]- encephalography
Single-Channel EEG Sensor
Cantero (EEG) signal IR devices HCI

47
 HUMAN COMPUTER machine
Mrs. Srilatha INTERACTION BASED HEAD learning with
Puli CONTROLLED MOUSE Python webcam HCI

Mouse Navigation Control Using Perceptual User robot

Aniket Kudale Head Motion Interface control HCI

Mouse Cusor Control Using Facial convolutional

Sashidhar R Movements - An HCI Application neural network webcam HCI

A Portable Wireless Head Movement

Controlled Human-Computer
Saurabh Barve Interface for People with Disabilities IR Sensor webcam HCI

Yashaswita
bhoir Hands-free Cursor Control OpenCV webcam HCI

Cursor Control using Head Tracking

Jonathan Ray and System Actions using Facial camera-frame
Bernal Castano Gestures extraction webcam HCI

COMPARATIVE STUDY OF
HAAR AND ADABOOST
ALGORITHMS FOR CURSOR
CONTROL USING HEAD AND
Anurag Singh EYE MOVEMENT FOR DISABLES OpenCV webcam HCI

Magnetic
Performance Analysis of a Head and
Angular Rate
Eye Motion-Based Control Interface
Gravity MARG-
for Assistive Robots
Sarah Stalljann (MARG)-sensor sensor HCI

Fuzzy based Computing for Disabled

Pradeep V using 3D Head Movement Fuzzification webcam HCI

Mouse Cursor Control Hands Free SVM and HOG

G. Abhishek Using Deep Learning algorithms SD card HCI

Virtual Hands: Real Time Keyboard,

Desktop & Application Navigation
Harshit Sharma using Gestures Open CV webcam HCI

Chern-Sheng A novel device for head gesture hidden Markov CCD

Lin measurement system in combination models camera HCI

48
 with eye-controlled human–machine (HMMs)
interface

HEAD-MOUSE: A SIMPLE
CURSOR CONTROLLER
BASED ON OPTICAL brainwave
Ali heydarigorji MEASUREMENT OF HEAD TILT (EEG) sensors IR LED HCI

VI. CONCLUSION AND FUTURE WORK

,accessible and user can use effectively.
Future Enhancements:
There is a need of a budget friendly , more
To further enhance the system’s capabilities and
accessible, and free hand gaze-tracking system
expand its potential applications, several
marks a great step toward improving human-
improvements can be made in future iterations:
computer interaction for individuals with
mobility impairments and for broader
applications such as gaming, virtual reality, and Future Enhancements:
assistive technology. By improving traditional
processing techniques we can enhance this
To further enhance the system’s capabilities and
system effectively detects and tracks movements
expand its potential applications, several
without relying on complex deep learning models
improvements can be made in future iterations:
budget friendly, privacy, and ease of
implementation.
Through real-time head tracking and blink 1. AI-Based Self-Learning for Automatic
detection, the system provides an intuitive way Calibration:
for users to interact with a PC, reducing the
dependency on conventional input devices like
o Using machine learning
keyboards and mice.
The One of the main advantages of our gaze- techniques to operate the system
tracking system is that it is more affordable and to automatically adjust interaction
accessible. Not like other advertised based on user behavior.
commercialized eye-tracking devices, which
often require specialized hardware and extensive o Minimizing the time limit
calibration, this system provides a viable required for traditional interaction
alternative using readily available components. and improve user-based accuracy.
Furthermore, the focus one real-time processing
and user adaptability ensures a flexible user
experience users with respect to different 2. Integration of Voice Commands &
environments. Gesture Recognition:

Despite its successes, certain limitations remain. o Making a hybrid interaction

The system's accuracy depends on lighting system by incorporating gaze
conditions and camera resolution, which may tracking and speech recognition.
affect tracking precision. In addition to this, head
movements beyond a certain limit can affect gaze
estimation. Noting these challenges are crucial o User-Friendly system that can
for making the system even more Strong take out maximum inputs by

49
 collaborating voice commands Features:
and eye movements.
o By adding more features like text-
3. Enhanced Cursor Movement & to-speech conversion which
Response Time: support users who have very
limited movement.
o By mean of modifying the
algorithms which are used to track o Creating user profiles that can be
the cursor that makes smoother changed to fit the needs of each
and more flexible movement. user and modifying interaction
settings accordingly.
o Reducing delay by improving
frame processing speed and VII. References
making enhance gaze-to-screen
mapping techniques.
[1]. Sancheti, K., Suhaas, A., & Suresh, P.
(2018, October). Hands-free cursor
4. Expansion to Smartphones & Smart control using intuitive head movements
Home Devices: and cheek muscle twitches. In TENCON
2018-2018 IEEE Region 10
Conference (pp. 0356-0361). IEEE.
o Adapting the system for mobile
devices to support hands-free
navigation on smartphones and [2]. Veeresh, V., Ajay, P., Varma, N. B.,
tablets. Anusha, J., & Divija, L. Human
Computer Interaction based Head
Controlled Mouse.
o Integrating with smart home
systems to enable eye-controlled
automation of household [3]. Rahmaniar, W., Ma’Arif, A., & Lin, T. L.
appliances. (2022). Touchless head-control (thc):
Head gesture recognition for cursor and
orientation control. IEEE Transactions
5. Improved Environmental
on Neural Systems and Rehabilitation
Adaptability:
Engineering, 30, 1817-1828.

o Improved Adaptability to the

[4]. Nangare, V., Samant, U., Rabha, S., &
Environment: o Improving the
Rodge, P. R. (2014). Controlling mouse
system's robustness so that it can
motions using eye blinks, head
function effectively in a variety of
movements and voice
lighting conditions (such as dim
recognition. computer, 3.
environments or outdoor
settings).

o Implementing adaptive
thresholding techniques to [5]. Tanusha, G., Havirbhavi, P., & Ashwini,
dynamically adjust contrast levels K. (2024, July). Enabling Cursor Control
for better eye detection. Through Eye Movement Using Hidden
Markov Model. In International
Conference on Data Science and Network
6. Support for Additional Accessibility

50
 Engineering (pp. 81-94). Singapore: analysis of a head and eye motion-based
Springer Nature Singapore. control interface for assistive
robots. Sensors, 20(24), 7162.
[6]. Molina-Cantero, A. J., Castro-García, J.
A., Gómez-Bravo, F., López-Ahumada, [12]. Pradeep, V. (2025). Fuzzy based
R., Jiménez-Naharro, R., & Berrazueta- Computing for Disabled using 3D Head
Alvarado, S. (2021). Controlling a mouse Movement. Informatica, 49(15).
pointer with a single-channel eeg
sensor. Sensors, 21(16), 5481.
[13]. Sharma, H., Saxena, R., Kumar,
S., Saini, A., Saad, M., & Faraz, M.
[7]. Kudale, A., Laddha, V., Thakare, R., & (2022, November). Virtual Hands: Real
Deshpande, L. (2012). Mouse Navigation Time Keyboard, Desktop & Application
Control Using Head Navigation using Gestures. In 2022
Motion. International International Conference on Fourth
Industrial Revolution Based Technology
and Practices (ICFIRTP) (pp. 262-267).
[8]. Journal of Advanced Research in
IEEE.
Computer Science, 3(3).

[14]. Lin, C. S., Ho, C. W., Chang, K.

[9]. Shashidhar, R., Snehith, K., Abhishek, P.
C., Hung, S. S., Shei, H. J., & Yeh, M. S.
K., Abhishek, B. V., & Pavitha, N. (2022,
(2006). A novel device for head gesture
April). Mouse Cusor Control Using
measurement system in combination with
Facial Movements-An HCI Application.
eye-controlled human–machine
In 2022 International Conference on
interface. Optics and Lasers in
Sustainable Computing and Data
Engineering, 44(6), 597-614.
Communication Systems (ICSCDS) (pp.
367-371). IEEE.
[15]. HeydariGorji, A., Safavi, S. M.,
Lee, C. T., & Chou, P. H. (2017,
[10]. Bernal Castano, J. R., Ota, A.,
December). Head-mouse: A simple
Soto, E. A., Kersting, B., Dietz, A. L.,
cursor controller based on optical
Yan, Q., ... & Domergue, G. (2025).
measurement of head tilt. In 2017 IEEE
Cursor Control using Head Tracking and
Signal Processing in Medicine and
System Actions using Facial Gestures.
Biology Symposium (SPMB) (pp. 1-5).
IEEE.
[11]. Stalljann, S., Wöhle, L., Schäfer,
J., & Gebhard, M. (2020). Performance

51
ORIGINALITY REPORT

4 %
SIMILARITY
2%
INTERNET SOURCES
0%
PUBLICATIONS
3%
STUDENT PAPERS
INDEX

PRIMARY SOURCES

Submitted to Meerut Institute of Engineering & Technology

1
Student Paper 2%
[Link]
2 Internet Source

Submitted to Liverpool John Moores University

<1%
3
Student Paper

[Link]
<1%
4 Internet Source

[Link]
5 Internet Source
<1%
[Link]
6 Internet Source
<1%
[Link]
7 Internet Source

<1%
<1%

52
53
5/9/25, 4:09 PM Meerut Institute of Engineering & Technology Mail - Abstract Acceptance Notification for Paper Id:ICACIT2025/RCSY/012

KUNAL YADAV <[Link].2021@[Link]>

Abstract Acceptance Notification for Paper Id:ICACIT2025/RCSY/012

1 message

ICACIT2025 <icacit2025titbpl@[Link]> Fri, May 9, 2025 at 4:01 PM

To: [Link].2021@[Link]

2nd International Conference on Advancements in Computational Intelligence

Technologies (ICACIT2025)
Bentham Science Publisher Double Indexed (Scopus & Web of Science)*

Abstract Acceptance Notification

Dear Author,

Last Date of Camera ready Paper Submission on or before: 31st May 2025
Conference Date: 20-21 June 2025

Paper ID: ICACIT2025/RCSY/012

Title: Eye and Head Movement-based Cursor Control System: A Survey

It is our pleasure to inform you that your abstract identified above has been ACCEPTED for further
processing in the proceeding of International Conference on Advancements in Computational
Intelligence Technologies (ICACIT2025). Therefore, you are required to submit your camera ready
paper with payment of Rupees 7500/- in the Bank account given below:

Account Name Technocrats Institute of Technology, Advance

Bank Name ICICI Bank
Branch Indrapuri Bhopal
Account NO. 028301001818
IFSC Code ICIC0001188

on or before 31st May 2025 along with Copyright Form attached herewith on google form link
given below-: [Link]

Website Link: [Link]

-----------------------------------------
With Warm Regards
Dr. Bhagwati Charan Patel
Editor-in-Chief,
Contact no: +91-9752578501
*Only Quality Papers
Disclaimer
Prospective authors should note that only original and previously unpublished research papers will be considered. Furthermore,
simultaneous submissions (under consideration for publication elsewhere) are not acceptable.

[Link] 1/1