Table 2 Mean (and standard deviation) for the different studied models considering the control image to generate the feature vectors
From: DecoFungi: a web application for automatic characterisation of dye decolorisation in fungal strains
Network | ERT | KNN | LR | MLP | RF | SVM |
|---|---|---|---|---|---|---|
DenseNet | 96.2(2.3) | 85.5(4.3) | 94.3(2.8) | 62.2(18.6) | 93.9(2.9) | 42.5(4.6) |
GoogleNet | 92.5(3.1) | 88.6(2.5) | 92.4(2.8) | 92.0(3.1) | 88.6(4.2) | 95.4(2.2) |
Inception v3 | 93.0(2.7) | 87.6(3.1) | 95.5(1.6) | 94.3(2.1) | 86.8(2.0) | 46.4(4.8) |
OverFeat | 87.2(2.4) | 82.6(4.5) | 92.7(2.1) | 92.2(2.6) | 82.0(3.5) | 93.0(2.4) |
Resnet 50 | 92.6(2.8) | 90.1(3.2) | 95.2(2.3) | 94.7(2.3) | 89.6(1.8) | 96.5(1.6) |
VGG16 | 95.0(2.0) | 86.4(2.3) | 94.7(1.7) | 92.4(1.7) | 89.2(3.5) | 33.1(4.3) |
VGG19 | 94.4(1.6) | 84.5(2.9) | 94.6(2.3) | 92.4(2.7) | 87.1(2.3) | 33.7(4.4) |
Xception v1 | 93.5(2.7) | 89.9(4.4) | 95.2(2.3) | 94.8(1.7) | 86.8(3.0) | 94.8(1.9) |