Table 8 Specificity and sensitivity analysis of deep learning models on SAVEE dataset.
From: Stacked convolutional neural network for emotion recognition using multi feature speech analysis
Sensitivity | angry | 0.92 | 0.91 | 0.56 | 0.84 | 0.41 | 0.91 | 0.88 | 0.5 | 0.9 | 0.42 | 0.95 | 0.94 | 0.53 | 0.78 | 0.33 | 0.92 | 0.92 | 0.53 | 0.82 | 0.39 |
disgust | 0.86 | 0.91 | 0.86 | 0.86 | 0.31 | 0.91 | 0.89 | 0.42 | 0.92 | 0.5 | 0.93 | 0.96 | 0.21 | 0.88 | 0.21 | 0.89 | 0.95 | 0.21 | 0.93 | 0.21 | |
fearful | 0.9 | 0.88 | 0.12 | 0.95 | 0.44 | 0.73 | 0.87 | 0.17 | 0.9 | 0.46 | 0.84 | 0.84 | 0.16 | 0.82 | 0.37 | 0.92 | 0.85 | 0.13 | 0.9 | 0.52 | |
happy | 0.84 | 0.9 | 0 | 0.88 | 0.6 | 0.86 | 0.86 | 0.12 | 0.9 | 0.57 | 0.88 | 0.9 | 0 | 0.84 | 0.62 | 0.88 | 0.9 | 0.7 | 0.84 | 0.59 | |
neutral | 0.98 | 0.98 | 0.15 | 0.96 | 0.79 | 0.98 | 0.91 | 0.42 | 0.94 | 0.58 | 0.97 | 0.94 | 0.39 | 0.97 | 0.77 | 0.96 | 0.97 | 0.52 | 0.97 | 0.79 | |
sad | 0.95 | 0.84 | 0.6 | 0.84 | 0.69 | 0.94 | 0.94 | 0.18 | 0.97 | 0.71 | 0.95 | 0.9 | 0.26 | 0.93 | 0.7 | 0.95 | 0.97 | 0.24 | 0.97 | 0.72 | |
surprised | 0.96 | 0.76 | 0.42 | 0.85 | 0.23 | 0.8 | 0.84 | 0.31 | 0.96 | 0.47 | 0.84 | 0.83 | 0 | 0.78 | 0.2 | 0.91 | 0.84 | 0.62 | 0.83 | 0.16 | |
Specificity | angry | 0.98 | 0.99 | 0.68 | 0.99 | 0.88 | 0.98 | 0.99 | 0.85 | 0.99 | 0.9 | 0.99 | 0.98 | 0.69 | 0.98 | 0.9 | 0.98 | 98 | 0.68 | 0.98 | 0.89 |
disgust | 0.99 | 0.99 | 0.91 | 0.98 | 0.94 | 0.98 | 0.98 | 0.87 | 0.99 | 0.93 | 0.99 | 0.99 | 0.9 | 0.98 | 0.94 | 0.99 | 0.99 | 0.92 | 0.98 | 0.84 | |
fearful | 0.98 | 0.97 | 0.9 | 0.98 | 0.94 | 0.97 | 0.97 | 0.9 | 1 | 0.97 | 0.97 | 0.98 | 0.89 | 0.96 | 0.9 | 0.97 | 0.98 | 0.92 | 0.99 | 0.88 | |
happy | 1 | 0.97 | 1 | 0.96 | 0.9 | 0.97 | 0.98 | 0.92 | 0.99 | 0.91 | 0.97 | 0.98 | 0.99 | 0.98 | 0.88 | 0.96 | 0.98 | 0.96 | 0.98 | 0.9 | |
neutral | 0.99 | 0.98 | 0.96 | 0.97 | 0.94 | 1 | 0.99 | 0.86 | 0.98 | 0.97 | 1 | 0.98 | 0.82 | 0.98 | 0.92 | 1 | 0.99 | 0.8 | 0.99 | 0.92 | |
sad | 0.99 | 0.99 | 0.96 | 0.99 | 0.9 | 0.99 | 0.98 | 0.91 | 1 | 0.94 | 1 | 0.99 | 0.81 | 0.99 | 0.88 | 0.99 | 0.99 | 0.86 | 0.99 | 0.9 | |
surprised | 0.97 | 0.98 | 0.68 | 0.99 | 0.92 | 0.96 | 0.98 | 0.89 | 0.97 | 0.86 | 0.97 | 0.98 | 1 | 0.97 | 0.96 | 0.99 | 0.98 | 0.98 | 0.97 | 0.97 | |
Feature Sets | Combined (196) | MFCC (40) | LPC (13) | Mel Spectrogram (128) | Chroma & Others (15) | Combined (196) | MFCC (40) | LPC (13) | Mel Spectrogram (128) | Chroma & Others (15) | Combined (196) | MFCC (40) | LPC (13) | Mel Spectrogram (128) | Chroma & Others (15) | Combined (196) | MFCC (40) | LPC (13) | Mel Spectrogram (128) | Chroma & Others (15) | |
Deep Learning | CNN + LSTM | CNN | LSTM | RNN + LSTM | |||||||||||||||||
Methods | |||||||||||||||||||||
