Research Article
A Novel Time-Incremental End-to-End Shared Neural Network with Attention-Based Feature Fusion for Multiclass Motor Imagery Recognition
Table 1
The details of the network model.
| Layers | Type |
| Conv 1 | Convolution 1 × 5 | Norm 1 | Batch normalization | Pool 1 | Max pooling 1 × 2 | Conv 2 | Convolution 1 × 3 | Norm 2 | Batch normalization | Pool 2 | Max pooling 1 × 2 | Drop 2 | Dropout layer | Conv 3 | Convolution 1 × 3 | Norm 3 | Batch normalization | Drop 3 | Dropout layer | Conv 4 | Convolution 1 × 3 | Norm 4 | Batch normalization | Flattened | Flattened layer | LSTM 1 | BiLSTM layer 32 | LSTM 2 | BiLSTM layer 20 | Fc 5 | Fully connected layer | Drop 5 | Dropout layer | Attention | Attention layer | Fc 6 | Fully connected layer | Classification | Softmax layer |
|
|