Figure 2.
The chart shows F 1 score values for bored, confused, engaged, frustrated, happy and neutral across four methods including text BERT, audio wav2vec 2, face ResNet 50 and multimodal fusion. For bored, scores are about 0.78 for text, 0.74 for audio, 0.76 for face and 0.83 for fusion. For confused, values are about 0.72, 0.69, 0.70 and 0.77. For engaged, values are about 0.81, 0.76, 0.79 and 0.85. For frustrated, values are about 0.73, 0.70, 0.71 and 0.79. For happy, values are about 0.85, 0.80, 0.83 and 0.88. For neutral, values are about 0.76, 0.72, 0.75 and 0.81. Multimodal fusion shows the highest scores across all emotional states.Emotion classification by modality