摘要
人脸识别是身份验证中最为关键的一步,因此基于MATLAB软件对人脸识别的各个流程及具体实现方法进行了研究。通过对所获取的图像进行灰度、均衡化以及中值滤波去噪预处理,去除图像中存在的干扰和噪声,得到更为清晰的图像之后基于肤色对待识别对象的脸部进行定位,同时利用灰度积分提取其眼睛部位,将测试图像导入GUI界面,与系统中提前录入的图像进行直方图对比,在匹配成功后显示最终的匹配识别结果。程序仿真结果表明,所研究的方法可以很好地完成各部分功能,提高系统识别的准确程度,达到有效识别人脸的目的。
Face recognition is the most critical step in identity verification,so the various processes and specific implementation methods of face recognition are studied based on MATLAB software.The obtained image is pretreated by grayscale,equalization and median filtering to remove the interference and noise in the image.After obtaining a clearer image,the face of the object to be recognized is located based on the skin color.At the same time,the eye part is extracted by grayscale integration.The test image is imported into the GUI interface and compared with the histogram of the image entered in advance in the system,the final matching recognition result is displayed after the matching is successful.The simulation results show that the proposed method can complete all functions well,improve the accuracy of system recognition,and achieve the purpose of effective face recognition.
作者
张丽芳
ZHANG Lifang(Shandong Huayu University of Technology,Dezhou 253034,China)
出处
《现代信息科技》
2023年第5期68-72,77,共6页
Modern Information Technology
基金
2021年山东华宇工学院科技计划项目(2021KJ12)
2021年度教学改革研究项目(2021JG01)。
