:::

網路暨嵌入式系統實驗室Networked & Embedded Systems Laboratory (NESL)指導教授：李昭賢 副教授 © 2023

:::

Wearable/Embedded Devices based on Deep Learning for Healthcare (以深度學習為基礎之穿戴/嵌入式健康照護裝置)

According to our observation on long-term healthcare industry and sports technologies, plantar pressure can be utilized for walking gait, motion analysis, and identity identification. Regarding identity identification, it further has the characteristic of privacy protection compared with the existing fingerprint identification and face identification. Based on the aforementioned assumptions, NESL focuses on (1) plantar pressure sensing and (2) intelligent identity recognition. First, we use the commercial pressure sensing module with the Bluetooth transmission mechanism to design and implement a software-defined optimization based on the characteristics of different people's plantar pressures. Second, the collected plantar pressures can be utilized to further complete the functional application of identity recognition through our proposed lightweight deep learning models.

基於過往創新穿戴式裝置之研發，以及過往長期健康照護產業之調查，或是運動科技之進展，足底壓力可用於行走步態、運動分析以及身份識別之用，其中，在身份識別部分，對比現行指紋識別、人臉識別，更具有隱私保護之特點。基於上述假設，本實驗室(NESL)將重點著重於(1)足底壓力感測以及(2)身份智慧識別。首先，我們運用市售之壓力感測片模組，並搭配藍芽傳輸機制，設計與實作出可依據不同人足底壓力之特徵，以軟體定義方式最佳化其壓力感測之功效表現；其次，感測的足底壓力可透過我們設計之輕量化深度學習模型，進一步完成身份識別之功能應用。

參考著作：

  • C. -H. Lee and H. -L. Wu, "Software-Defined Design of In-shoe Plantar Pressure Measurement Mechanism," Proceedings of IEEE International Conference on Consumer Electronics (ICCE), Las Vegas, NV, USA, Jan. 6-8, 2023
  • C. -H. Lee and C. -H. Lee, "Lightweight Improvement of Deep Learning Model on Plantar Pressure Images," Proceedings of IEEE 11th Global Conference on Consumer Electronics (GCCE), Osaka, Japan, Oct. 18-21, 2022, pp. 611-612
  • C. -H. Lee and L. -T. Li, "Cost-Effective Person Identity Recognition based on Plantar Pressure Images," Proceedings of IEEE International Conference on Consumer Electronics - Taiwan (ICCE-Taiwan), Taoyuan, Taiwan, Sep. 28-30, 2020

開放式原始碼：

回頁首