International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 03 | Mar 2024
p-ISSN: 2395-0072
www.irjet.net
Sign Language Recognizer Smita Wagh1, Diksha Jaiswal2, Snehal Thalkar3, Ekata More4, Resha Deshmukh5 1Guide and Professor, Department of Computer Engineering 2,3,4,5 Students, Department of Computer Engineering
Jayawantrao Sawant College of Engineering, Pune, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Sign languages serve as vital tools for
Sign Language (ASL), French Sign Language (FSL), and Indian Sign Language (ISL), the latter being our primary focus. India hosts a significant population of hearingimpaired individuals, with ISL being the preferred mode of communication for over a million deaf adults and approximately half a million deaf children. Despite its prevalence, research on ISL lags behind that of ASL. Thus, our research endeavors to bridge this gap and facilitate effective communication between hearing and speechimpaired individuals and the broader community.
facilitating communication within the deaf and hard-ofhearing community, enabling them to interact effectively with hearing individuals. While extensive research has been conducted in American Sign Language (ASL), Indian Sign Language (ISL) has received comparatively less attention from researchers worldwide. One of the primary challenges faced in the advancement of ISL recognition systems is the lack of standardized datasets and the significant linguistic variation across different regions. To address these challenges, we present a novel approach for Indian Sign Language (ISL) gesture recognition, focusing on both singlehanded and two-handed gestures. Unlike existing systems that often require signers to wear gloves or use marker devices for hand segmentation, our proposed method eliminates such requirements, thus simplifying the recognition process. Our approach leverages Convolutional Neural Networks (CNNs) for image classification, offering improved accuracy and robustness in ISL gesture recognition. By utilizing a custom-built dataset comprising continuous ISL gestures captured using a laptop webcam in home environments, we aim to enhance the accessibility and usability of ISL recognition technology.
In addition to overcoming linguistic barriers, our Sign Language recognition system also addresses the technological challenges associated with such implementations. We employ advanced computer vision and machine learning techniques to recognize and interpret Sign Language gestures accurately. By leveraging modern technologies, our system aims to achieve real-time recognition, enhancing the efficiency and accessibility of communication for hearing-impaired individuals. Furthermore, our research extends beyond mere recognition to include the development of educational tools and resources for learning Sign Language. By creating interactive tutorials and applications, we seek to empower both hearing-impaired individuals and the general public to engage with Sign Language more effectively, thereby fostering inclusivity and understanding within society.
Key Words: Sign Language, ISL, CNN, Deep Learning, Deaf and Dumb.
1. INTRODUCTION
2. LITERATURE SURVEY
Humans employ diverse means of communication, including verbal speech in various regional languages and non-verbal expressions. Sign Language, specifically tailored for the Deaf and Hard of Hearing, serves as their primary mode of communication, utilizing gestures to convey messages without relying on speech. Unlike spoken languages, Sign Languages cannot be transcribed into written form. However, the challenge arises when attempting to bridge the communication gap between those proficient in Sign Language and those unfamiliar with it. To address this issue, we propose a Sign Language recognition system.
Deaf people, who live in villages usually, do not have access to sign language. However, in all large towns and cities across the Indian subcontinent, deaf people use sign language which is not standard sign language. Extensive work and awareness program are being done for implementation of ISL in education systems. Zaw Hein and Thet Paing Htoo [1] worked on skin color-based enhancement method and color-based segmentation method for detecting skin color of hands, and manual signs of Myanmar Sign Language Recognition System based on machine learning. They proposed, sign classification following horizontal and vertical projection employs Support Vector Machine (SVM), utilizing Gaussian radial basic function for accurate classification, leveraging SVM's capabilities in supervised machine learning for
Our system aims to mitigate the communication barrier by facilitating communication between hearingimpaired individuals and the general population. Sign Language exhibits variations across different countries, each with its own vocabulary and grammar, such as American
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 666