Scholarlink

<p>Current fashion design applications combining augmented reality (AR) and smart wearables are limited by low technical integration, poor interactive experience, and lack of clothing customization, which restricts users’ immersion in interactive clothing experiences. This article therefore adopts a multimodal sensor data fusion and real-time AR (augmented reality) rendering method based on long short term memory (LSTM) network to integrate the IMU (Inertial Measurement Unit) posture data, EMG (electromyography) signals, and physiological parameters from optical sensors into LSTM model after time synchronization and normalization to extract the temporal features of human motion and physiological state, and to use the parameter mapping to drive the AR rendering engine to adjust the material texture, geometric deformation, and lighting effects in real time to achieve millisecond level dynamic clothing mapping and interaction. Experimental results show that the method can achieve average latency of 10.9ms for highly complex actions, 99.5% rendering stability for long duration and detailed manipulation scenes, and clothing response latency of 125ms for electrodermal activity (EDA) at 4&#xa0;s interval, which verifies the effectiveness of the method in enhancing interactivity in fashion design and provides a technical solution for deep integration of augmented reality and smart wearable technology.</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

Interactivity of fashion design combining augmented reality (AR) and smart wearables

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月16日

查看全文
  • Qinyu Bai,
  • Yufan Wen,
  • Jiabai Jin,
  • Jiaji Zhou

摘要

Current fashion design applications combining augmented reality (AR) and smart wearables are limited by low technical integration, poor interactive experience, and lack of clothing customization, which restricts users’ immersion in interactive clothing experiences. This article therefore adopts a multimodal sensor data fusion and real-time AR (augmented reality) rendering method based on long short term memory (LSTM) network to integrate the IMU (Inertial Measurement Unit) posture data, EMG (electromyography) signals, and physiological parameters from optical sensors into LSTM model after time synchronization and normalization to extract the temporal features of human motion and physiological state, and to use the parameter mapping to drive the AR rendering engine to adjust the material texture, geometric deformation, and lighting effects in real time to achieve millisecond level dynamic clothing mapping and interaction. Experimental results show that the method can achieve average latency of 10.9ms for highly complex actions, 99.5% rendering stability for long duration and detailed manipulation scenes, and clothing response latency of 125ms for electrodermal activity (EDA) at 4 s interval, which verifies the effectiveness of the method in enhancing interactivity in fashion design and provides a technical solution for deep integration of augmented reality and smart wearable technology.