基于改进稠密网络的视频监控人脸识别算法研究OA
Research on Video Surveillance Face Recognition Algorithm Based on Improved Dense Network
为了提升视频监控中的人脸识别能力,研究利用运动历史图像算法来实现人体跟踪,提出了一种改进稠密网络.在结果中显示,研究采用的人体跟踪算法的跟踪准确率高达 99.5%,同时提出的识别算法的识别准确率能够稳定在 99.7%以上,且能够针对不同表情特征的人脸表现出较高的识别准确率.以上结果表明,改进稠密网络能够有效提升视频监控人脸识别能力,对城市监控的智能化发展具有重要意义.
In order to improve the ability of face recognition in video surveillance,it studies the use of motion history image algorithm to realize human tracking,and proposes an improved dense network.The results show that the tracking accuracy of the human tracking algorithm used in the study is as high as 99.5%,and the recognition accuracy of the proposed recognition algorithm can be above 99.7%,and can show high recognition accuracy for faces with different expression features.The above results show that the improved dense network can effectively improve the face recognition ability of video surveillance,which is of great significance to the intelligent development of urban surveillance.
余鸣
曲靖职业技术学院 信息技术系,云南 曲靖 655000
计算机与自动化
视频监控运动历史图像算法改进稠密网络人体跟踪人脸识别
video surveillancemotion history image algorithmimproved dense networkhuman trackingface recognition
《现代信息科技》 2024 (001)
89-93 / 5
评论