CAJ | 학술논문

通过对现有人脸识别方法的分析,提出了基于离散小波变换和离散余弦变换相结合的图像特征提取算法,和基于显式分解策略的SVM分类器的设计方法。设计了一套基于人脸识别的门禁系统,给出了硬件结构和软件工作流程。经现场测试,该系统运行稳定,对人脸图像的识别率较高,能满足实际应用的各种要求。
통과대현유인검식별방법적분석,제출료기우리산소파변환화리산여현변환상결합적도상특정제취산법,화기우현식분해책략적SVM분류기적설계방법。설계료일투기우인검식별적문금계통,급출료경건결구화연건공작류정。경현장측시,해계통운행은정,대인검도상적식별솔교고,능만족실제응용적각충요구。
Through the analysis of existing face recognition method,the image feature extraction algorithm base on discrete wavelet transform and discrete cosine transform combined was proposed,and the design method of SVM classifier base on explicit decomposition strategy was given.An entrance guard system based on face recognition was designed.The hardware structure and software workflow was given.The system ran stably though field testing.The recognition rate of face image was higher.The system can meet the various requirements of practical application.