An intelligent dimming method based on human expression and gesture fusion recognition for smart home

Haoyun Li, Kun Gao*, Shuhao Zhang, Shuo Wang, Ruitao Ding, Yunfan Bao

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

According to the demand of future smart home for lighting experience, an embedded intelligent LED dimming system based on the Raspberry Pi 4B is proposed. The system detects the human proximity distance and captures the gesture movements and other parameters through one ADPS-9960 sensor. An embedded miniature CSI camera OV5647 is used to collect the user's facial expression image data. PWM signal is generated from Raspberry Pi GPIO pin to control the LED illumination. PWM duty cycle range is determined by the fusion results of normalized weighted modulation from the expression type recognized from the facial images and the gesture detected by ADPS-9960. Under the basic 3-tier CNN recognition model framework, the expression recognition is further enhanced by introducing the fusion with Face landmarks and HOG features. The experimental results show that the proposed dimming system can adaptively control the luminaire brightness according to the user's psychological state and behavior, and can light off when the person leaves the light, which meets the requirements of modern smart home for user light experience and energy saving.

源语言英语
主期刊名Ninth Symposium on Novel Photoelectronic Detection Technology and Applications
编辑Junhao Chu, Wenqing Liu, Hongxing Xu
出版商SPIE
ISBN(电子版)9781510664432
DOI
出版状态已出版 - 2023
活动9th Symposium on Novel Photoelectronic Detection Technology and Applications - Hefei, 中国
期限: 21 4月 202323 4月 2023

出版系列

姓名Proceedings of SPIE - The International Society for Optical Engineering
12617
ISSN(印刷版)0277-786X
ISSN(电子版)1996-756X

会议

会议9th Symposium on Novel Photoelectronic Detection Technology and Applications
国家/地区中国
Hefei
时期21/04/2323/04/23

指纹

探究 'An intelligent dimming method based on human expression and gesture fusion recognition for smart home' 的科研主题。它们共同构成独一无二的指纹。

引用此