一种基于改进VGG16网络的人脸表情识别算法

A Facial Expression Recognition Algorithm Based on Improved VGG16 Network

针对神经网络对人脸表情进行识别时,使用的分类损失函数主要是交叉熵损失函数,导致网络对于不同人脸表情类别识别率不高的问题,将类别注意力机制和上下文感知金字塔引入VGG16网络,产生类别损失函数,与交叉熵损失函数一起作为网络训练的损失函数,从而提高网络的人脸表情识别准确率。实验结果表明:改进后的VGG16网络在人脸表情数据集RAF-DB和FERPLUS上有比原始VGG16网络具有更高的人脸表情识别率。

When the neural network recognizes facial expressions,the classification loss function used is mainly the cross-en-tropy loss function,which leads to the problem that the network has a low recognition rate for different facial expression categories.The category attention mechanism and context awareness pyramid are introduced into the VGG16 network to generate a category loss function,which together with the cross-entropy loss function is used as the loss function for network training,so as to improve the accuracy of facial expression recognition of the network.The experimental results show that the improved VGG16 network has a high-er facial expression recognition rate on the facial expression datasets RAF-DB and FERPLUS than the original VGG16 network.