Skip to main content
. 2022 Nov 3;10:958870. doi: 10.3389/fpubh.2022.958870

Table 7.

Comparison of experimental results with different kernel sizes.

Kernel size Trained accuracy Tested accuracy
Layer 2 Layer 1
1 × 1 7 × 7 0.7216 0.6948
5 × 5 0.7523 0.7090
3 × 3 0.8669 0.6818
3 × 3 7 × 7 0.7207 0.6870
5 × 5 0.7505 0.6896
3 × 3 0.8594 0.6688
5 × 5 7 × 7 0.7216 0.7210
5 × 5 0.7495 0.7013

The bold values are the best values obtained in the experiment.