ANJUMAN POLYTECHNIC
SADAR, NAGPUR
ACADEMIC YEAR 2024-25
DEPARTMENT OF COMPUTER ENGINEERING
PROJECT
SUBJECT: CPP
TOPIC: Sign language to text
converter
NAME ROLL NO
Karamjeet Singh Khokhar 24
Mohammad Zeeshan 31
Mohammad Khamgavwala 32
Mubasshir Khan 33
Under the guidance of: - Jahangir Ansari Sir
Acknowledgments: -
I would like to express my sincere gratitude to everyone
who will provide their guidance and support during this
project, "Sign Language to Text Converter."
First and foremost, I would like to thank my guide,
[Jahangir Sir], for their continuous support, expert
guidance, and valuable feedback throughout the course
of this project. Their mentorship will be instrumental in
ensuring the successful completion of this project.
Additionally, I would like to acknowledge the
contributions of researchers and developers in the field
of sign language recognition and assistive
technologies, whose work has inspired and laid the
foundation for this project’s development.
Abstract: -
This project aims to develop a Sign Language to Text
Converter to help bridge the communication gap
between sign language users and non-sign language
speakers. Using computer vision and machine learning
techniques, the system will recognize hand gestures and
convert them into text in real time. By capturing hand
and finger movements with a camera, the software will
identify sign language gestures and display
corresponding text. This tool will enhance accessibility
and inclusivity for individuals who use sign language,
allowing them to interact more effectively in daily life.
The project will explore machine learning models and
image recognition algorithms to create an efficient,
accurate, and user-friendly solution. Ultimately, this
system aims to improve communication and support
inclusivity, making it easier for sign language users to
engage in various platforms and devices.
Content page:
1.Introduction and background of the industry or user
based problem.
2. Literature survey for problem identification and
specification.
3. Proposed detailed methodology of solving the
identified problem with action plan.
1. Introduction and background of the
industry or user based problem: -
1.1 Introduction to the Project
This project aims to develop a Sign Language to Text Converter to
bridge the communication gap between sign language users and those
unfamiliar with it. The system will leverage computer vision and
machine learning to recognize sign language gestures and convert them
into readable text in real-time. By providing a tool for seamless interaction
between sign language users and non-users, this project aims to improve
accessibility and inclusivity, particularly in social and professional
settings.
1.2 Background of the Industry
The assistive technology industry has advanced significantly, with
solutions such as speech-to-text and hearing aids improving
communication for people with disabilities. However, sign language
remains a major communication barrier for the deaf and hard-of-hearing
community, especially when interacting with people who do not know
sign language. While existing solutions, like video interpreters and
translation apps, are available, they often face limitations in accuracy,
speed, and accessibility. This project seeks to address these issues by
developing a real-time, machine learning-based sign language recognition
system that is both efficient and accessible.
1.3 User-Based Problem
For many deaf and hard-of-hearing individuals, sign language is the
primary mode of communication. Unfortunately, not everyone
understands sign language, leading to social and communication barriers.
This problem is compounded by the lack of accurate, real-time translation
systems. Current tools often fail to provide reliable and instant results,
limiting their effectiveness. This project will focus on creating a system
that can efficiently translate common sign language gestures into text,
improving communication and fostering inclusivity in everyday
interactions.
1.4 Scope of the Project
This project will focus on translating Sign Language gestures into text
using machine learning techniques. The system will initially cover a
limited set of common gestures, with potential for future expansion. The
project will focus on creating a user-friendly interface, and performance
will be evaluated based on speed, accuracy, and usability.
2. Literature survey for problem
identification and specification: -
2.1 Overview of Sign Language Recognition
Sign language recognition is an evolving field that aims to
translate sign language gestures into text or speech. Early
systems relied on sensor-based devices, like gloves, but
these were often cumbersome. With advances in computer
vision and machine learning, vision-based systems that use
cameras to capture gestures have become more popular.
These systems analyse images or videos in real-time to
identify hand gestures and translate them into text.
2.2 Existing Technologies
Vision-based systems: These use CNNs and camera
feeds to track and recognize gestures in real time.
They are non-invasive but can struggle with accuracy
and speed.
Sensor-based systems: Gloves with sensors offer
precise tracking but require the user to wear additional
equipment.
Mobile applications: Apps that use smartphone
cameras are more accessible, but many lack real-time
processing or struggle with complex gestures.
2.3 Problem Identification and Specification
This project addresses the limitations of existing
systems, focusing on improving accuracy, real-time
processing, and accessibility. The goal is to develop a
Sign Language to Text Converter that can:
Recognize a wide range of gestures accurately.
Process gestures in real-time with minimal delay.
Be accessible and easy to use for both sign language
users and non-users.
3. Proposed detailed methodology of solving
the identified problem with action plan: -
3.1 System Overview
The objective of this project is to develop a real-time Sign Language to
Text Converter that translates sign language gestures into text using
machine learning and computer vision techniques. The system will
capture real-time video input, process the video frames to identify
gestures, and output the corresponding text. This solution aims to facilitate
better communication between sign language users and non-sign language
speakers.
3.2 Methodology
i. Data Collection
A dataset of American Sign Language (ASL) gestures
will be collected. This will include commonly used
gestures (e.g., alphabets, numbers, and basic words).
Existing datasets or custom video recordings of
gestures will be used.
ii. Preprocessing
The collected data will undergo image preprocessing to
normalize brightness, resize the images, and apply
augmentation techniques (such as flipping, rotation) to
enhance model generalization.
iii. Feature Extraction
Key features such as hand shape, position, and motion
will be extracted from the video frames.
Techniques like Convolutional Neural Networks (CNNs)
will be used to capture the features from images.
iv. Model Development
A CNN or other deep learning model will be trained to
classify hand gestures into the corresponding sign
language characters or words.
The model will be trained on the preprocessed dataset
using techniques like backpropagation and gradient
descent to minimize classification errors.
v. Real-Time Gesture Recognition
The trained model will be deployed in a real-time
system where it processes live camera feed.
The system will recognize gestures and display
corresponding text on the screen.
vi. Post-Processing
Post-processing techniques will be applied to refine the
output text, ensuring it is contextually accurate,
especially for complex or ambiguous gestures.
3.3 Tools and Technologies
Programming Language: HTML,CSS, JavaScript,
Bootstrap, Node js
Libraries: OpenCV (for video capture),
TensorFlow/Keras (for machine learning), NumPy (for
numerical operations)
Hardware: Webcam
Platform: Desktop application
References and Bibliography: -
References
No external references were used directly in the
development of this
Bibliography
No additional resources were consulted during
the project.
