首页|多目标人脸识别系统的设计与实现

多目标人脸识别系统的设计与实现

扫码查看
多目标人脸识别系统是指能够同时识别多个目标个体的人脸识别系统,但面临识别速度与准确性平衡、姿态光照变化适应性等问题.为此,本文设计了一个集成YOLOv3、支持向量机(Support Vector Machine,SVM)与Dlib的人脸识别系统.系统流程涵盖人脸图像采集、检测算法设计及识别实现.利用Dlib的高效检测器完成人脸定位与关键点提取,经预处理后,YOLOv3负责关键区域识别,SVM则用于特征训练与分类.实验结果表明,系统在保持高识别准确率的同时,在多目标场景中表现优异.
Design and implementation of multi-target face recognition system
Multi-target face recognition system is a face recognition system that can recognize multiple target individu-als at the same time,but it faces challenges in balancing recognition speed and accuracy,as well as adapting to varia-tions in pose and lighting.Therefore,this paper introduces a face recognition system integrating YOLOv3,Support Vector Machines(SVM)and Dlib.The system process includes face image acquisition,detection algorithm design,and recognition implementation.Dlib's efficient detector is used to complete face location and key point extraction.After preprocessing,YOLOv3 is responsible for key region identification,and SVM is used for feature training and classification.The experimental results demoenstrate that the system maintains high recognition accuracy and performs excellently in multi-target scenarios.

multi-target face recognitionYOLOv3SVM

丁苍峰、任亮、刘洁

展开 >

延安大学 数学与计算机科学学院,陕西 延安 716000

多目标人脸识别 YOLOv3 SVM

国家自然科学基金项目国家级大学生创新创业训练计划项目延安大学自然科学基金项目延安大学教学改革研究项目

62262067202210719034YDBK2018-35YDJG23-27

2024

延安大学学报(自然科学版)
延安大学

延安大学学报(自然科学版)

影响因子:0.238
ISSN:1004-602X
年,卷(期):2024.43(3)
  • 7