Conference paper Open Access

Through-Screen Visible Light Sensing Empowered by Embedded Deep Learning

Liu, Hao; Ye, Hanting; Yang, Jie; Wang, Qing


Citation Style Language JSON Export

{
  "DOI": "10.1145/3485730.3493454", 
  "title": "Through-Screen Visible Light Sensing Empowered by Embedded Deep Learning", 
  "issued": {
    "date-parts": [
      [
        2021, 
        11, 
        5
      ]
    ]
  }, 
  "abstract": "<p>Motivated by the trend of realizing full screens on devices such as smartphones, in this work we propose through-screen sensing with visible light for the application of fingertip air-writing. The system can recognize handwritten digits with under-screen photodiodes as the receiver. The key idea is to recognize the weak light reflected by the finger when the finger writes the digits on top of a screen. The proposed air-writing system has immunity to scene changes because it has a fixed screen light source. However, the screen is a double-edged sword as both a signal source and a noise source. We propose a data preprocessing method to reduce the interference of the screen as a noise source. We design an embedded deep learning model, a customized model ConvRNN, to model the spatial and temporal patterns in the dynamic and weak reflected signal for air-writing digits recognition. The evaluation results show that our through-screen fingertip air-writing system with visible light can achieve accuracy up to 91%. Results further show that the size of the customized ConvRNN model can be reduced by 94% with less<br>\nthan a 10% drop in performance.</p>", 
  "author": [
    {
      "family": "Liu, Hao"
    }, 
    {
      "family": "Ye, Hanting"
    }, 
    {
      "family": "Yang, Jie"
    }, 
    {
      "family": "Wang, Qing"
    }
  ], 
  "id": "5646942", 
  "event-place": "Coimbra, Portugal", 
  "type": "paper-conference", 
  "event": "Workshop on Workshop on Challenges in Artificial Intelligence and Machine Learning for Internet of Things in conjuction with ACM SenSys 2021 (AIChallengeIoT)"
}
24
11
views
downloads
Views 24
Downloads 11
Data volume 14.3 MB
Unique views 24
Unique downloads 11

Share

Cite as