Table 2 in Identification of morphologically cryptic species with computer vision models: wall lizards (Squamata: Lacertidae: Podarcis) as a case study

Table 2. Evaluation of the three tested architectures in the four datasets for the two-class experiment. Models with the highest accuracy are highlighted in bold
SexViewMetric*Inception V3ResNet 50Inception ResNetV2Combined predictionsMalesDorsalAccuracy0.9350.9220.9480.955AUC0.9760.9820.9840.982F1 Pboc0.9510.9410.9620.967F1 Plus0.9050.8870.9190.930Head lateralAccuracy0.9260.9290.9360.930AUC0.9720.9750.9840.976F1 Pboc0.9460.9470.9530.949F1 Plus0.8820.8950.8890.885FemalesDorsalAccuracy0.8730.9060.9050.943AUC0.9310.9650.9700.970F1 Pboc0.9050.9300.9290.948F1 Plus0.8100.8570.8590.908Head lateralAccuracy0.9350.9190.9270.911AUC0.9620.9590.9760.972F1 Pboc0.9530.9410.9470.937F1 Plus0.8970.870.8720.847
*AUC refers to the area under the ROC curve. Podarcis bocagei was the positive class. F1, harmonic mean of precision and recall; Pboc, Podarcis bocagei; Plus, Podarcis lusitanicus.