Table 2 Comparisons of mAPs on Oxford Flowers-17 under FGIR.

From: Learning optimal image representations through noise injection for fine-grained search

Method

Splits

Mean

1

2

3

4

5

LBP59

0.098

0.099

0.101

0.103

0.102

0.101

HOG58

0.111

0.113

0.112

0.111

0.115

0.112

ResNet18 (Pretrained)

0.512

0.509

0.513

0.515

0.518

0.513

Yang et al.21 (Vgg-16)

-

-

-

-

-

0.877

Kumar et al.29

0.901

0.923

0.946

0.931

0.940

0.928

Our Method

0.947

0.934

0.959

0.939

0.947

0.946

  1. Significant values are in [bold].
Back to article page