Audio-Textual Emotion Recognition Based on Improved Neural Networks

Figure 3

The structure of the Bi-LSTM network for textual feature learning.