Table 3.

Comparison with other SOTA methods.

Model	Number of subjects	Accuracy
DNN (Li et al., 2018)	95	85.3%
Combined MCNNEs (Aghdam et al., 2019)	459	70.45%
CNN-EW (Xing et al., 2018)	1096	66.88%
ASD-DiagNet (Eslami et al., 2019)	1035	70.1%
cGCN (Wang et al., 2021)	1057	70.7%
3D CNN (Thomas et al., 2020)	1162	64%
	95	87.38 %
	459	77.42 %
ID-GCN(our model)	867	76.3 %
	1066	71.44 %