Skip to main content

Table 2 Evaluation of Ffirst on available leaf datasets: Austrian Federal Forests, Flavia, Foliage, Swedish, Middle European Woods and Leafsnap

From: Fine-grained recognition of plants from images

 

AFF

Flavia \(10\times 40\)

Flavia \(\frac{1}{2}\times \frac{1}{2}\)

Foliage

Swedish

MEW

Leafsnap

Leafsnap top 5

Num. of classes

5

32

32

60

15

153

185

185

\(\hbox {Ffirst}_\text {a}^{\forall +}\) (1)

\(97.1\pm 1.5\)

\(99.4\pm 0.3\)

\(99.2\pm 0.2\)

99.2

\(99.7\pm 0.3\)

\(98.8\pm 0.2\)

\(81.2\pm 1.8\)

\(95.9\pm 1.5\)

\(\hbox {Ffirst}_\text {i}^{\forall +}\) (2)

\(97.3\pm 1.6\)

\(99.3\pm 0.3\)

\(98.9\pm 0.3\)

98.1

\(99.7\pm 0.3\)

\(98.4\pm 0.2\)

\(73.1\pm 2.3\)

\(92.4\pm 1.7\)

\(\hbox {Ffirst}_\text {b}^{\forall +}\) (3)

\(99.5\pm 0.6\)

\(99.3\pm 0.4\)

\(99.0\pm 0.2\)

98.3

\(99.4\pm 0.5\)

\(97.9\pm 0.2\)

\(77.2\pm 1.9\)

\(94.8\pm 1.5\)

\(\hbox {Ffirst}_{ib\sum }^{\forall +}\) (4)

\(100.0\pm 0.0\)

\(99.7\pm 0.3\)

\(99.6\pm 0.1\)

99.3

\(99.8\pm 0.2\)

\(99.3\pm 0.1\)

\(81.8\pm 1.2\)

\(96.5\pm 1.1\)

\(Ffirst _{ib\prod }^{\forall +}\) (5)

\(100.0 \pm 0.0\)

\(99.8 \pm 0.3\)

99.7 ± 0.1

99.3

\(99.8 \pm 0.3\)

\(99.5 \pm 0.1\)

\(83.7 \pm 1.1\)

\(97.3 \pm 1.1\)

Inception-ResNet-v2 +maxout

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

99.9+

\(-\)

\(-\)

Kumar et al. [12]

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

\(\approx\) 73

96.8

Fiel, Sablatnig [11]

93.6

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

Novotný, Suk [22]

\(-\)

\(-\)

91.5

\(-\)

\(-\)

84.9

\(-\)

\(-\)

Karuna et al. [23]

\(-\)

\(-\)

96.5

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

Kadir et al. [18]

\(-\)

95.0

\(-\)

95.8

\(-\)

\(-\)

\(-\)

\(-\)

Lee et al. [21]

\(-\)

97.2

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

\(-\)

Qi et al. [27]

\(-\)

\(-\)

\(-\)

\(-\)

99.4

\(-\)

\(-\)

\(-\)