基于核DCV算法的主动近红外人脸识别方法

Kernel Discriminative Common Vector Method for Active NIR Face Recognition

针对人脸识别领域存在的受环境光照变化影响大的问题,分析了各种主动成像方法的特点以及人脸皮肤的光谱反射特性,提出了使用近红外LED灯作为主动光源,选用近红外滤光片配合CCD相机完成人脸图像采集,并综合运用可鉴别共同向量方法(DCV)和核投影方法进行人脸特征提取。该特征提取方法同时解决了核投影方法面临的大样本问题和可鉴别共同向量方法面临的样本维数较高问题,减少了计算复杂性,提高了特征提取的效率和准确度。仿真结果表明基于核DCV的主动近红外人脸识别方法有利于消除光照影响、提高识别效率。

Aiming at reducing the influence on face recognition of varying illumination, a new face recognition method is proposed after analyzing the characteristics of various active imaging methods and the spectral reflection characteristic of human skin. The new method uses active near-infrared imaging way to acquire human face images and uses the combination of kernel projection and discriminative common vectors（DCV） to achieve facial feature extraction for recognition. So the problems of large sample size and high dimensions can be solved together, which means it can reduce the computation complexity and improve the efficiency and accuracy of feature extraction. The results of experiments prove the effectiveness of proposed method.