Table 5.

Classification results on the dataset split on the subject level.

Method	Data Type	Number of Parameters	FLOPS	Mean Accuracy (%)	STD	Best Accuracy (%)
Pre-ResNet with DS	O2Hb	0.7 M	15 M	87.25	0.472	88.07
Pre-ResNet with Stage 1, (1 × 5)		0.7 M	15 M	90.14	0.649	91.28
EfficientNetV2 m with DS [55]		52 M	225 M	91.05	0.562	91.97
PyramidNet with DS [51]		17 M	396 M	87.39	0.481	88.89
CF-CNN with DS [52]		29.7 M	627 M	89.27	0.531	90.74