Table 9.
VMMR results on VMMRDB-495 [4].
Method | 1B | 2B–2S | ||||||
---|---|---|---|---|---|---|---|---|
Classification Level | Make | Model | Make | Model | ||||
Model Recogniser Backbone | AlexNet [15] | Acc (%) | 70.81 | 60.69 | 96.20 | 72.26 | ||
F1 (%) | 63.64 | 53.63 | 95.40 | 66.87 | ||||
G (%) | 1.66 | 3.92 | 98.34 | 96.08 | ||||
ResNet50 [16] | Acc (%) | 97.24 | 92.38 | 97.55 | 92.55 | |||
F1 (%) | 97.12 | 89.50 | 97.47 | 89.74 | ||||
G (%) | 20.80 | 20.80 | 79.20 | 67.24 | ||||
DenseNet201 [17] | Acc (%) | 98.03 | 93.60 | 98.24 | 93.70 | |||
F1 (%) | 97.92 | 91.04 | 98.16 | 91.11 | ||||
G (%) | 23.25 | 34.89 | 76.75 | 65.11 |