Structure of the network module
| Module name | Functional | Network architecture |
|---|---|---|
| BiLSTM | Local feature extraction | Conv1d(k = 3, s = 1, p = 1), BatchNorm1d(64), ReLU() |
| MaxPool1d() | ||
| Conv1d(k = 3, s = 1, p = 1), BatchNorm1d(128), ReLU() | ||
| MaxPool1d() | ||
| Contextual relationship | BiLSTM(hidden = 128(Video)/256(Vibrate)) | |
| Focus on important features | Multi-head Attention(num_heads = 4) | |
| GCN | Spatial feature | GCNConv(hidden = 128(Video)/256(Vibrate)), ReLU() |
| GCNConv(), ReLU() | ||
| GCNConv(), ReLU(), Global Average Pooling() | ||
| Characteristic fusion | Multisource spatial-temporal characterization | Concat() |
| Linear() | ||
| Dropout(0.3) | ||
| Linear() |
| Module name | Functional | Network architecture |
|---|---|---|
| BiLSTM | Local feature extraction | Conv1d(k = 3, s = 1, |
| MaxPool1d() | ||
| Conv1d(k = 3, s = 1, | ||
| MaxPool1d() | ||
| Contextual relationship | BiLSTM(hidden = 128(Video)/256(Vibrate)) | |
| Focus on important features | Multi-head Attention(num_heads = 4) | |
| GCN | Spatial feature | GCNConv(hidden = 128(Video)/256(Vibrate)), ReLU() |
| GCNConv(), ReLU() | ||
| GCNConv(), ReLU(), Global Average Pooling() | ||
| Characteristic fusion | Multisource spatial-temporal characterization | Concat() |
| Linear() | ||
| Dropout(0.3) | ||
| Linear() |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.