CNN-LSTM Facial Expression Recognition Method Fused with Two-Layer Attention Mechanism

Figure 1

CNN-LSTM model the structure of global feature fusing attention mechanism. In CNN local feature extraction layer, C means the convolutional layer and S means the pooling layer. The elements in {} represent the size and the number of the convolution kernels, respectively. The convolutional layers and the pooling layers have step size of 1 and 2, respectively.