基于多视角解耦融合的多模态会话情感识别

张婷婷,朱小飞

Multi-view decoupled fusion for multimodal emotion recognition in conversation

Tingting ZHANG,Xiaofei ZHU

表 3 不同模型在MELD数据集上的性能对比结果

Tab.3 Performance comparison of different models on MELD dataset

模型 F1% Acc% W-F1% 中性 惊讶 悲伤 快乐 愤怒 DialogueRNN# 76.97 47.69 20.41 50.92 45.52 60.31 57.66 MMGCN# 76.33 48.15 26.74 53.02 46.09 60.42 58.31 CTNet# 77.40 52.70 32.50 56.00 44.60 62.00 60.50 MM-DFN$ 77.76 50.69 22.93 54.78 47.82 62.49 59.46 SCMM$ — — — — — — 59.44 CMCF-SRNet$ — — — — — 62.80 62.30 MultiDAG$ — — — — — 64.41 64.00 FrameERC$ 80.00 58.10 42.54 63.37 56.27 67.62 66.33 AdaIGN$ 79.75 60.53 43.70 64.54 56.15 67.62 66.79 MDF 79.86 59.03 44.84* 64.60 57.63* 68.06* 66.92*