基于LBP与卷积神经网络的人脸识别

Face recognition based on LBP and convolutional neural network

卷积神经网络在人脸识别上有较好的效果,但是其提取的人脸特征忽略了人脸的局部结构特征.为了提取更加全面的人脸特征,提出一种基于局部二值模式(local binary pattern,LBP)与卷积神经网络相结合的新方法.首先,提取人脸图片的LBP特征图像,然后把LBP图像与原RGB图像结合作为网络输入数据,并且使用随机梯度下降法训练网络参数,最后用训练得到的网络模型对人脸图片进行识别.通过在LFW(labeled face in the wild)人脸识别数据库上的实验表明,在卷积神经网络中加入LBP图像信息可以提高人脸识别的准确率.另外,当增加训练数据时,提出的方法得到的识别率会进一步提高,更说明提出方法的有效性.

Convolutional neural network （CNN） has achieved promising results in face recognition, but it ignores the local structure features of faces when it extracts face features. In order to extract more comprehensive face feature, a new method based on local binary pattern （LBP） and convolutional neural network was presented. At first, LBP feature maps were extracted from face images. Then,the combination of LBP images and RGB images were input to convolutional neural network, and optimize network paramenters using stochastic gradient descent method. At last, the trained CNN model was used to recognize unseen face samples. On LFW face database, the experimental results show that the proposed method has a better recognition performance. In addition, when more training sets were added, the face recognition was higher than before, which showed the effectivity of that method.