Scholarlink

<p>Emotion-aware computing is becoming increasingly significant, especially in patient–robot interactions, where recognizing and responding to emotions is critical for effective communication. However, current systems struggle with accuracy and real-time performance when processing multimodal emotional input such as facial expressions, speech, and physiological signs. This study tackles these issues by introducing a hybrid deep learning model for emotion-aware patient–robot interactions. The model uses convolutional neural networks for spatial feature extraction and long short-term memory networks for sequential pattern recognition, together with an attention mechanism to improve multimodal data fusion. This guarantees that important emotional cues are accurately captured and processed. The suggested approach beats current models with 92.5% accuracy, 93.2% precision, and 91.8% recall. It also has faster training times and a smaller model size, making it ideal for real-time use in patient–robot interaction scenarios. This work contributes to emotion-aware computing by providing an effective way for improving human–robot interaction in healthcare and beyond.</p>

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

Modality-aware hybrid deep learning framework with attention for emotion-aware patient–robot interaction

  • 期刊论文

  • 发表日期：

    2026年6月15日

查看全文
  • Rajanikanth Kashi Nagaraj,
  • H H Surendra,
  • H R Archana

摘要

Emotion-aware computing is becoming increasingly significant, especially in patient–robot interactions, where recognizing and responding to emotions is critical for effective communication. However, current systems struggle with accuracy and real-time performance when processing multimodal emotional input such as facial expressions, speech, and physiological signs. This study tackles these issues by introducing a hybrid deep learning model for emotion-aware patient–robot interactions. The model uses convolutional neural networks for spatial feature extraction and long short-term memory networks for sequential pattern recognition, together with an attention mechanism to improve multimodal data fusion. This guarantees that important emotional cues are accurately captured and processed. The suggested approach beats current models with 92.5% accuracy, 93.2% precision, and 91.8% recall. It also has faster training times and a smaller model size, making it ideal for real-time use in patient–robot interaction scenarios. This work contributes to emotion-aware computing by providing an effective way for improving human–robot interaction in healthcare and beyond.