Scholarlink

Hand gestures are a primary medium for people with hearing disabilities to communicate verbally with each other. In light of this problem, several technologies were developed to bridge the gap between deaf and hearing communities by making software systems. Systems that can automatically process the identification of sign language gestures will help solve this ongoing problem. The reviewed studies were published from 2015 to 2024 and retrieved from trusted journal databases, including Science Direct and IEEE Explore. A total of 15 pertinent articles were then finalized and analyzed. The analysis includes its strengths and limitations as there are growing publications on hand gesture recognition using computer vision, which implemented Convolutional Neural Networks as its primary technique in sign language recognition. In terms of performance, the recognition accuracy ranges from 84.68% to 99.13% in the chosen studies. Though most models present high accuracy, their potential for possible deployment is dependent on the technique used. The primary factor influencing the potential deployment of sign language recognition in real-world settings is CNN’s heavy demand for computing resources. Addressing these factors will be crucial for the widespread adoption and deployment of CNN-based recognition systems. The path of this research is promising, and further progress will be expected if challenges are resolved.

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 Recognition Using Convolutional Neural Networks (CNN)—A Literature Review

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Shannon John G. Cantos,
  • Joshua Mark A. Magwili,
  • Jonna L. Simporios,
  • Cereneo S. Santiago,
  • Erwin L. Cahapin,
  • Beverly A. Malabag

摘要

Hand gestures are a primary medium for people with hearing disabilities to communicate verbally with each other. In light of this problem, several technologies were developed to bridge the gap between deaf and hearing communities by making software systems. Systems that can automatically process the identification of sign language gestures will help solve this ongoing problem. The reviewed studies were published from 2015 to 2024 and retrieved from trusted journal databases, including Science Direct and IEEE Explore. A total of 15 pertinent articles were then finalized and analyzed. The analysis includes its strengths and limitations as there are growing publications on hand gesture recognition using computer vision, which implemented Convolutional Neural Networks as its primary technique in sign language recognition. In terms of performance, the recognition accuracy ranges from 84.68% to 99.13% in the chosen studies. Though most models present high accuracy, their potential for possible deployment is dependent on the technique used. The primary factor influencing the potential deployment of sign language recognition in real-world settings is CNN’s heavy demand for computing resources. Addressing these factors will be crucial for the widespread adoption and deployment of CNN-based recognition systems. The path of this research is promising, and further progress will be expected if challenges are resolved.