Research Article
Facial Expression Recognition Algorithm Based on Fusion of Transformed Multilevel Features and Improved Weighted Voting SVM
Table 2
Multilayer convolutional network structure.
| | Layer | Input (WHD) | Kernel_num | Kernel_size | Stride | Pad | Out (WHD) |
| | Conv1 | 14141024 | 1024 | 3 | 1 | 0 | 12121024 | | Conv2 | 12121024 | 1024 | 3 | 1 | 0 | 10101024 | | Conv3 | 10101024 | 1024 | 3 | 1 | 0 | 881024 | | Conv4 | 881024 | 1024 | 3 | 1 | 0 | 661024 | | Conv5 | 661024 | 1024 | 3 | 1 | 0 | 441024 | | Conv5 | 441024 | 1024 | 3 | 1 | 0 | 221024 | | Conv6 | 221024 | 512 | 2 | 1 | 0 | 11512 | | Conv7 | 11512 | 256 | 1 | 1 | 0 | 11256 | | Conv8 | 11256 | 8 | 1 | 1 | 0 | 118 |
|
|
