本课题研究的步骤如下：先提取人脸的特征向量；产生训练样本和测试样本；再用LVQ创建神经网络模型，该模型用训练样本进行训练调整权值；用测试样本对建立的人脸朝向识别模型进行验证，要求有较高的识别率。

本课题要求使用LVQ神经网络的算法进行Matlab仿真，对人脸朝向进行有效的判断和识别。

-This study is the following steps: first extract facial feature vector　generate training and testing samples　reuse create LVQ neural network model, which is trained using training samples to adjust the weights　using the test sample towards the establishment of a human face recognition model validated requires a higher recognition rate. This topic requires the use of LVQ neural network algorithm Matlab simulation, the human face towards effective judgment and identification.