Table 6 Performance of the three models under the Modulation Spectrum Smoothing attack (UA/WA/ACC)
From: Black-box adversarial attacks through speech distortion for speech emotion recognition
Model | VTLN | McAdams | MSS | ALL |
|---|---|---|---|---|
CNN-LSTM | 55.57/57.21/56.39 | 57.84/58.49/58.17 | 56.04/57.39/56.72 | 52.74/55.93/54.34 |
GCN | 66.73/67.38/67.06 | 67.32/69.47/68.40 | 66.94/67.30/67.12 | 61.93/62.29/62.11 |
CNN-MAA | 67.32/69.19/68.26 | 66.04/68.84/67.44 | 66.73/69.72/68.23 | 63.83/65.37/64.60 |