Research on a Microexpression Recognition Technology Based on Multimodal Fusion

Figure 11

(a) Accuracy of CNN + LSTM under CASME II, (b) loss function of the training model.

(a)

(b)