|
基于视频分析的伪装人脸检测系统的设计与实现 |
Design and implementation of a disguised face detection system based on video analysis |
投稿时间:2018-12-24 修订日期:2018-12-24 |
DOI: |
中文关键词: 伪装人脸检测,背景差分法,帧间差分法,筛选式级联分类器 |
英文关键词: Disguised face detection, Background difference method, Inter-frame difference method, Screening cascade classifier |
基金项目:2016年安徽省高校科学研究项目(项目编号:KJ2016A764,KJ2016A757);2017年安徽省高校科学研究项目(项目编号:KJ2017A560,KJ2017A562);2017年高校优秀青年人才支持计划项目--(项目编号:gxyqZD2017141);安徽省级质量工程大学生创客实验室(项目编号:2016ckjh224);2018年安徽省高校科学研究项目(项目编号:KJ2018A06970)校级科技创新团队(项目编号:Wzykj2018A02) |
|
摘要点击次数: 40 |
全文下载次数: 0 |
中文摘要: |
伪装人脸检测为伪装人脸的跟踪、人脸识别等其他应用提供基础,本文设计了基于视频分析的伪装人脸检测系统,包含用户检测模块,墨镜伪装检测模块,口罩伪装检测模块和帽子伪装检测模块。提出一种基于视频分析的伪装检测方法,该方法首先通过背景差分法和连续帧间差分法进行人脸定位,然后采用基于Haar特征的筛选式级联分类器进行伪装检测。算法简单,时间效率高,易于实现,提高了检测效率,适合ATM机的硬件环境。实验表明,该方法可取得较高的检测率,具有良好的应用价值。 |
英文摘要: |
Disguised face detection can provide the foundation for other applications such as the tracking of disguised faces and human recognition. This paper designs a camouflage face detection system based on video analysis, which includes the detection module, the sunglasses camouflage detection module, the mask camouflage detection module and the hat camouflage detection module.A disguised detection method based on video analysis is proposed. The method is based on the background difference method and the continuous frame difference method to locate the face first, and then the screening cascade classifier based on the Haar feature is used to detect the face camouflage. The algorithm is simple, efficient, easy to implement, and improves the detection efficiency. It is suitable for the hardware environment of ATM machine. The results show that the method can achieve high detection rate and has good application value. |
View Fulltext
查看/发表评论 下载PDF阅读器 |
关闭 |