Research Article
Inception-LSTM Human Motion Recognition with Channel Attention Mechanism
Table 2
Experimental results of different models on PAMAP2.
| Algorithm model | Accuracy | Precision | Recall | F1 value | Model size |
| Classics CNN | 91.11% | 91.41% | 91.11% | 91.26% | 7.14 M | LSTM | 89.28% | 89.69% | 89.28% | 89.49% | 0.37 M | Neural network without ECA | 92.44% | 92.93% | 92.44% | 92.68% | 2.67 M | Original ECA | 93.91% | 94.04% | 93.91% | 93.97% | 2.68 M | This article model | 95.04% | 95.06% | 95.21% | 95.13% | 2.68 M | Literature: [24] | 89.30% | — | — | — | — | Literature: [25] | 92.97% | — | — | — | — | Literature: [19] | 93.16% | — | — | — | 3.51 M |
|
|