Research Article
Facial Expression Recognition of Instructor Using Deep Features and Extreme Learning Machine
Table 7
Accuracy comparison between convolution neural networks on standard datasets along with time taken per frame.
| CNN model | JAFFE (%) | CK (%) | FER2013 (%) | Time taken per frame (sec) | Number of parameters (million) |
| AlexNet [86] | 93 | 90.2 | 61.1 | — | 61 | VGG16 [86] | 96 | 92.4 | 59.6 | 0.94 | 138 | VGG19 [87] | 93 | 93 | 60 | — | 144 | ResNet101 [88] | 90 | — | 49 | — | 44.5 | Inceptionv3 [89] | 75.8 | 76.5 | — | — | 23 | Proposed method | 96.8 | 86.5 | 62.5 | 0.74 | 0.005 |
|
|