面向智能驾驶的人脸检测研究进展
Research Progress on Face Detection for Intelligent Driving
王丹 1洪杰 1钟亮洁 1雷春光 1许楠升1
作者信息
- 1. 肇庆学院机械与汽车工程学院 广东省肇庆市 526061
- 折叠
摘要
智能驾驶作为汽车领域的一项前沿技术,日益受到广泛关注.人脸检测在智能驾驶中扮演着重要的角色,通过实时监测驾驶员和乘客的脸部特征能够提高车辆内部环境感知能力,增强系统对驾驶员和乘客的理解,从而提升交通安全性、舒适性和个性化服务体验.文章首先回顾了基于传统特征提取的人脸检测算法,然后介绍了基于深度学习的几种主流检测方法,包括基于级联卷积网络、单阶段检测及双阶段检测算法,分析了这几种算法的结构和优缺点并介绍了轻量级检测方法.最后,对未来的研究方向进行了展望.
Abstract
As a cutting-edge technology in the automotive field,intelligent driving has attracted more and more attention.Face detection plays an important role in intelligent driving,and real-time monitoring of the driver's and passengers'facial features can improve the vehicle's internal environment perception and enhance the system's understanding of the driver and passengers,thereby improving traffic safety,comfort and personalized service experience.This paper first reviews the face detection algorithm based on traditional feature extraction,and then introduces several mainstream detection methods based on deep learning Xi,including cascaded convolutional network,single-stage detection and two-stage detection algorithm,analyzes the structure,advantages and disadvantages of these algorithms,and introduces lightweight detection methods.Finally,the future research directions are prospected.
关键词
人脸检测/深度学习/计算机视觉/车载嵌入式系统Key words
Face detection/Profpund study/Computer vision/Vehicle-mounted embedded system引用本文复制引用
出版年
2024
NETL
NSTL
万方数据