基于机器视觉的人脸生物特征改进识别技术

摘要

In the traditional identification method,the recognition rate of human face biological features is improved by means of spectral regression,matrix integrity constraint,and obtaining clean input data from the original data with noise,but the recognition interference brought by low rank projection matrix is ignored,which causes the problems of low recognition preci-sion and poor recognition efficiency. Therefore,a human face biological feature improved recognition technology based on ma-chine vision is proposed. Cameras are used as machine vision to collect human face images. The human face image dichromatic reflection model is constructed to perform preprocessing of human face images. On the basis of machine vision,the improvement of human face biological feature recognition is realized by combining with the LDBP model. The experimental results show that in comparison with the traditional identification technology,the improved recognition technology for human face biological fea-ture recognition has high recognition precision and strong practicability,which has certain advantages.%传统识别方法主要通过谱回归和矩阵完整性约束,从带噪声的原始数据中得到干净的输入数据,达到提高识别人脸生物特征的目的,但忽略了低秩投影矩阵对识别带来的干扰,导致识别精度低、效率差的问题,故提出基于机器视觉的人脸生物特征改进识别技术.通过摄像机代替机器视觉对人脸图像进行采集,构建人脸图像双色反射模型,对人脸图像进行预处理,并以机器视觉为基础,结合局部差分二值模型(LDBP),实现对人脸生物特征识别的改进.实验结果表明,相比于传统识别技术,采用改进识别技术在进行人脸生物特征识别方面的识别精度较高,实用性较强,具有一定的优势.