Scholarlink

It is estimated that by 2050 over 700 million people or one in every ten people will have disabling hearing loss [16] according to the World Health Organization (WHO). These individuals may face a range of communication challenges in their daily lives. These challenges can vary depending on the degree of hearing loss, the individual’s communication preferences, and the environment they are in. Bridging their communication with society is really a serious concern of today’s era. Through this study, the researchers propose a Sign Language Interpreter (SLI) mobile application using a machine learning framework viz TensorFlow lite, object detection model viz SSDMobinet. This application recognizes six American Sign Language (ASL) hand gestures, two English words and translates them into text and speech. A comparative study is performed with other applications (desktop and mobile based). Further, the application is deployed in the cloud using firebase to be used publicly. The proposed mobile application is evaluated in real life scenarios and found to have an accuracy of 74.44%. Though the accuracy is low, the aim of the work is to introduce a real time SLI that is portable, without IOT interfaces or devices and the communication between person without hearing loss and person with hearing loss is effective.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Sign Language Interpreter Using Machine Learning Model for Mobile Devices

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pinky Panda,
  • Bikram Ghosh,
  • Jyotshna Dongardive

摘要

It is estimated that by 2050 over 700 million people or one in every ten people will have disabling hearing loss [16] according to the World Health Organization (WHO). These individuals may face a range of communication challenges in their daily lives. These challenges can vary depending on the degree of hearing loss, the individual’s communication preferences, and the environment they are in. Bridging their communication with society is really a serious concern of today’s era. Through this study, the researchers propose a Sign Language Interpreter (SLI) mobile application using a machine learning framework viz TensorFlow lite, object detection model viz SSDMobinet. This application recognizes six American Sign Language (ASL) hand gestures, two English words and translates them into text and speech. A comparative study is performed with other applications (desktop and mobile based). Further, the application is deployed in the cloud using firebase to be used publicly. The proposed mobile application is evaluated in real life scenarios and found to have an accuracy of 74.44%. Though the accuracy is low, the aim of the work is to introduce a real time SLI that is portable, without IOT interfaces or devices and the communication between person without hearing loss and person with hearing loss is effective.