数据增强策略在人脸识别技术实现的研究

以小数据集为样本进行卷积神经网络模型的训练过程,容易出现所得到的神经网络模型泛化能力不足的问题。传统的处理方法大都通过数据增强的方式来提高训练数据的样本总数。本文选取多个网络模型进行对比实验,验证不同神经网络在训练过程中是否使用数据随机增强方式的模型识别准确率提高的效果,为如何选取小数据集样本训练神经网络提供参考。

Research on the Implementation of Data Enhancement Strategy in Face Recognition Technology

The process of training a convolutional neural network model with a small data set as a sample is prone to insufficient generalization ability of the neural network model obtained by training. In the face of this problem, most of the traditional processing methods are to increase the total number of training data samples through data enhancement to achieve the purpose of improving the accuracy of the neural network model. This paper selects multiple network models to conduct comparative experiments to verify whether different neural networks use random data enhancement methods to improve the accuracy of model recognition during the training process. Provide reference for the selection of training neural network with small data set samples.