Scholarlink

Hand gesture recognition is one of the most important applications in human-computer interaction, especially in improving accessibility for deaf and hard-of-hearing people. However, current solutions have drawbacks in real-time performance, user-independence, and accurate temporal pattern recognition. This paper presents an approach towards real-time dynamic gesture recognition both using the combination of LSTM networks and the data collection framework based on cameras. It captures gesture via camera, extracts key points of the hand using MediaPipe, and trains an LSTM model to classify gestures according to predefined actions. Therefore, it increases accuracy as well as allows for interaction to be applied in the communication through gesture.

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

Dynamic Gesture Recognition Using LSTM for Real-Time Indian Sign Language Prediction

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Anirudh Singh Rautela,
  • Esam Ashfaq,
  • Barish Priyam Chetia,
  • Ishaan Bhadrike,
  • Pranay Chauhan,
  • Dipti Theng,
  • Madhuri Hiwale

摘要

Hand gesture recognition is one of the most important applications in human-computer interaction, especially in improving accessibility for deaf and hard-of-hearing people. However, current solutions have drawbacks in real-time performance, user-independence, and accurate temporal pattern recognition. This paper presents an approach towards real-time dynamic gesture recognition both using the combination of LSTM networks and the data collection framework based on cameras. It captures gesture via camera, extracts key points of the hand using MediaPipe, and trains an LSTM model to classify gestures according to predefined actions. Therefore, it increases accuracy as well as allows for interaction to be applied in the communication through gesture.