基于人工智能技术的高光谱人脸自动化识别系统设计
Design of Hyperspectral Facial Automatic Recognition System Based on Artifi-cial Intelligence Technology
张绍龙1
作者信息
- 1. 西安职业技术学院 大数据应用学院,西安 710077
- 折叠
摘要
为有效识别人脸区域,提升其在多种领域应用效果,设计基于人工智能技术的高光谱人脸自动化识别系统.以模块化思想设计嵌入式系统架构,采集与预处理高光谱人脸图像,并将预处理后的图像数据放入RAM存储器;人脸检测模块调用RAM存储器存储数据,并加载Haar人脸分类器,完成人脸区域检测提取工作;之后由人脸特征提取与识别模块经人脸区域LBP特征提取、LeNet-5卷积神经网络人脸识别模型构建与训练等操作,输出人脸识别结果.实验结果表明,该系统能够在较短时间内完成LeNet-5卷积神经网络人脸识别模型训练.
Abstract
In order to recognize face region effectively and improve its application effect in various fields,a hyperspec-tral face automatic recognition system based on artificial intelligence technology is designed.The embedded system ar-chitecture is designed with modularity idea,and the hyperspectral face image is collected and preprocessed,and the preprocessed image data is put into RAM memory.Face detection module calls RAM memory to store data,and loads Haar face classifier to complete face region detection and extraction.After that,face recognition results are outputed by the face feature extraction and recognition module through LBP feature extraction of face region,LeNet-5 convolutional neural network face recognition model construction and training.The experimental results show that the system can complete the LeNet-5 convolutional neural network face recognition model training in a relatively short time.
关键词
人工智能技术/高光谱/自动化/人脸识别/LBP特征/LeNet-5网络Key words
artificial intelligence technology/hyperspectral/automation/face recognition/LBP features/LeNet-5 network引用本文复制引用
基金项目
西安职业技术学院2021年度科研骨干研究项目(2021GG05)
陕西省科技计划一般项目(2020NY-163)
出版年
2024
NETL
NSTL
维普
万方数据