Scholarlink

In order to automatically interpret sign language gestures and translate them into spoken or written English, Sign Language Recognition combines computer vision and natural language processing. The aim is to enable smooth communication between individuals who use sign language and those who do not by interpreting gestures, movements, postures, and facial expressions that correspond with elements of spoken language. As a conduit between their primary form of communication and the general public, this technical invention is very important to the deaf community. It becomes clear that the proposed strategy for Arabic sign language recognition is essential for promoting successful communication among deaf and hard-of-hearing people. Arabic sign language gestures are categorized using the RegNet, MobileNet, ResNet, ConvNeXt, and SqueezeNet models. Specifically, RegNet and ConvNeXt show better F1-scores, recall, and precision. The program predicts a hand motion in Arabic sign language and outputs the matching text, improving user accessibility to communication. To maximize its usefulness, the model is implemented on a Raspberry Pi to provide a compact size and enhanced portability.

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

Arabic Sign Language Recognition and Voice Translation Device Using Embedded Machine Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Asfak Ali,
  • Bibek Das,
  • Saifuddin Sk,
  • Singhan Ganguly,
  • Sheli Sinha Chaudhuri

摘要

In order to automatically interpret sign language gestures and translate them into spoken or written English, Sign Language Recognition combines computer vision and natural language processing. The aim is to enable smooth communication between individuals who use sign language and those who do not by interpreting gestures, movements, postures, and facial expressions that correspond with elements of spoken language. As a conduit between their primary form of communication and the general public, this technical invention is very important to the deaf community. It becomes clear that the proposed strategy for Arabic sign language recognition is essential for promoting successful communication among deaf and hard-of-hearing people. Arabic sign language gestures are categorized using the RegNet, MobileNet, ResNet, ConvNeXt, and SqueezeNet models. Specifically, RegNet and ConvNeXt show better F1-scores, recall, and precision. The program predicts a hand motion in Arabic sign language and outputs the matching text, improving user accessibility to communication. To maximize its usefulness, the model is implemented on a Raspberry Pi to provide a compact size and enhanced portability.