Three-dimensional face reconstruction using a single two-dimensional image is a hot topic in the field of image processing.Inspired by deep convolutional neural network（CNN）and 3 D morphable model（3 DMM）

a method of 3 D face reconstruction using convolution neural network to regress the shape and expression parameters of 3 DMM is proposed.Based on the convolution neural network model VGG-16

an improved network model of VGG-BN is designed.By adding batch normalization layer after each convolution layer

the performance of network model is optimized.The pre-training model is introduced into the training of VGG-BN network by transfer learning method.The improved network model is trained on 300 W-LP dataset

tested on AFLW2000-3 D dataset

and compared with existing methods.Experimental results show that the improved network model can improve the accuracy and generalization of face reconstruction

and the effect of face reconstruction is better.