Table 6. Rank-1 accuracy comparison of proposed method with other methods using two sessions of FV_USM dataset.

Train/Test split	Method	Rank-1 accuracy (%)
Training 4 Testing 2	Boucherit et al. (2022) *	96.15
	Tamang & Kim (2022) *	97.22
	FV-EffResNet*	97.56
Training 6 Testing 6	Boucherit et al. (2022) **	81.71
	Chai et al. (2022) **	94.67
	Das et al. (2018)	97.53
	FV-EffResNet**	95.75
FV-EffResNet	98.79
Training 8 Testing 4	Liu et al. (2022)	99.75
	FV-EffResNet	99.54
Training 10 Testing 2	Ma, Wang & Hu (2023)	99.90
	FV-EffResNet	99.39

Notes.

^*Training and Testing with only first session.

^**Training with first session and Testing with second session.

Results with no superscript indicate first and second session combined.