Table 4.

The performance results of different 2D and 3D CNN models on PET image.

Model	Acc	AUC	F1	Precision	Recall	AP
3D-CNN [14]	0.8482	0.8797	0.8095	0.8151	0.8041	0.7339
2D-CNN [17]	0.8241	0.8446	0.8069	0.7778	0.8383	0.7229
Jo et al. [15]	0.8451	0.8649	0.8162	0.8506	0.7844	0.7618
Hybrid [19]	0.8714	0.8654	0.8537	0.8512	0.8563	0.7919
Max.	0.8583	0.8787	0.8333	0.8599	0.8084	0.7791
Avg.	0.8661	0.8904	0.8440	0.8625	0.8263	0.7888
ARP [24]	0.8609	0.8811	0.8349	0.8701	0.8024	0.7848
LWP(current study)	0.8871	0.9088	0.8693	0.8827	0.8563	0.8189