abstract |
The present invention provides an emotion recognition method and device based on immersive virtual environment and multi-modal physiological signals. The method includes: step 1, conducting an emotion induction experiment in an experimental environment; step 2, sitting a trainee in the experimental environment On the smart mobile device, wear a virtual reality device to watch the pictures selected from the International Emotional Picture System, collect the trainees' biological signals, and use the SAM self-assessment body model to quantify the pictures; step 2, locate the quantitative score at valence‑ The arousal two-dimensional emotion model is mapped to determine the emotion type corresponding to the quantitative score; step 3, the physiological signal of the trainee is input into the recognition convolutional neural network model with the emotion type as the label, and the emotion model is obtained by training; step 4, obtain the tester's Personal biosignals are imported into the emotion model to identify the tester's emotion category. |