Real-Time Multi-Scale Face Detector on Embedded Devices
Zhao, Xu1,2; Liang, Xiaoqing1,2; Zhao, Chaoyang1,2; Tang, Ming1,2; Wang, Jinqiao1,2
发表期刊SENSORS
ISSN1424-8220
2019-05-01
卷号19期号:9页码:22
通讯作者Zhao, Chaoyang(chaoyang.zhao@nlpr.ia.ac.cn)
摘要Face detection is the basic step in video face analysis and has been studied for many years. However, achieving real-time performance on computation-resource-limited embedded devices still remains an open challenge. To address this problem, in this paper we propose a face detector, EagleEye, which shows a good trade-off between high accuracy and fast speed on the popular embedded device with low computation power (e.g., the Raspberry Pi 3b+). The EagleEye is designed to have low floating-point operations per second (FLOPS) as well as enough capacity, and its accuracy is further improved without adding too much FLOPS. Specifically, we design five strategies for building efficient face detectors with a good balance of accuracy and running speed. The first two strategies help to build a detector with low computation complexity and enough capacity. We use convolution factorization to change traditional convolutions into more sparse depth-wise convolutions to save computation costs and we use successive downsampling convolutions at the beginning of the face detection network. The latter three strategies significantly improve the accuracy of the light-weight detector without adding too much computation costs. We design an efficient context module to utilize context information to benefit the face detection. We also adopt information preserving activation function to increase the network capacity. Finally, we use focal loss to further improve the accuracy by handling the class imbalance problem better. Experiments show that the EagleEye outperforms the other face detectors with the same order of computation costs, on both runtime efficiency and accuracy.
关键词face detection ARM-based devices model acceleration computer vision
DOI10.3390/s19092158
收录类别SCI
语种英语
资助项目Natural Science Foundation of China[61772527] ; Natural Science Foundation of China[61806200] ; Natural Science Foundation of China[61876086] ; Natural Science Foundation of China[61772527] ; Natural Science Foundation of China[61806200] ; Natural Science Foundation of China[61876086]
项目资助者Natural Science Foundation of China
WOS研究方向Chemistry ; Electrochemistry ; Instruments & Instrumentation
WOS类目Chemistry, Analytical ; Electrochemistry ; Instruments & Instrumentation
WOS记录号WOS:000469766800202
出版者MDPI
七大方向——子方向分类AI芯片与智能计算
引用统计
被引频次:14[WOS]   [WOS记录]     [WOS相关记录]
文献类型期刊论文
条目标识符http://ir.ia.ac.cn/handle/173211/23713
专题模式识别国家重点实验室_图像与视频分析
通讯作者Zhao, Chaoyang
作者单位1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China
第一作者单位模式识别国家重点实验室
通讯作者单位模式识别国家重点实验室
推荐引用方式
GB/T 7714
Zhao, Xu,Liang, Xiaoqing,Zhao, Chaoyang,et al. Real-Time Multi-Scale Face Detector on Embedded Devices[J]. SENSORS,2019,19(9):22.
APA Zhao, Xu,Liang, Xiaoqing,Zhao, Chaoyang,Tang, Ming,&Wang, Jinqiao.(2019).Real-Time Multi-Scale Face Detector on Embedded Devices.SENSORS,19(9),22.
MLA Zhao, Xu,et al."Real-Time Multi-Scale Face Detector on Embedded Devices".SENSORS 19.9(2019):22.
条目包含的文件 下载所有文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
sensors-19-02158-v2.(3135KB)期刊论文作者接受稿开放获取CC BY-NC-SA浏览 下载
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[Zhao, Xu]的文章
[Liang, Xiaoqing]的文章
[Zhao, Chaoyang]的文章
百度学术
百度学术中相似的文章
[Zhao, Xu]的文章
[Liang, Xiaoqing]的文章
[Zhao, Chaoyang]的文章
必应学术
必应学术中相似的文章
[Zhao, Xu]的文章
[Liang, Xiaoqing]的文章
[Zhao, Chaoyang]的文章
相关权益政策
暂无数据
收藏/分享
文件名: sensors-19-02158-v2.pdf
格式: Adobe PDF
此文件暂不支持浏览
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。