基于BP神经网络的人脸识别系统设计

Design of Face Recognition System Based on BP Neural Network

以人脸识别为目标,将BP神经网络与子空间分析相结合,运用SVD分解的方法将特征值分解出来,再经过训练达到人脸识别的效果,最后得出待测人的有效身份。实验结果表明融合SVD分解方法获取特征值的BP神经网络可以有效提高识别精确率。

This article aims to achieve facial recognition by combining BP neural network with subspace analysis,using SVD decomposition method to decompose feature values,and then training to achieve the effect of facial recognition.Finally,the effective identity of t he tested person is obt ained.The experimental results show t hat the recognition accuracy of the BP neural network using the fused SVD decomposition method to obt ain feature values is improved.