Real-Time Human-Music Emotional Interaction Based on Deep Learning and Multimodal Sentiment Analysis

<div>The Extended Cohn-Kanade Dataset (a) and Bimodal Dataset (b). The CK+ dataset images have been preprocessed with face extraction, colour removal, and equal scale cutting. The weight occupied by the bimodal in decision-level fusion is determined by individually matching the emotion text with the color face image in the bimodal emotion dataset that I created.</div>

Wireless Communications and Mobile Computing

fig9

Figure 9

Figure 9: Real-Time Human-Music Emotional Interaction Based on Deep Learning and Multimodal Sentiment Analysis 