Deep Transfer Learning for Biology Cross-Domain Image Classification

Table 9

Accuracy (%) results of fine-tuning on ImageNet.

Model	Flowers17	Flowers102	Plant Seedling	PlanktonSet 1.0	QUT Fish
AlexNet	91.18	77.23	96.80	75.17	39.55
VGG-16	96.76	80.94	98.15	75.85	49.27
GoogLeNet-v3	97.06	93.06	98.19	76.49	59.15
ResNet-18	97.96	88.21	98.00	76.44	51.12
ResNet-34	96.47	88.73	98.19	76.55	52.54
ResNet-50	97.65	90.29	98.53	77.27	52.24
ResNet-101	97.06	90.55	98.44	77.48	52.78
ResNet-152	97.06	90.54	98.48	77.19	53.25

The entries in boldface indicate the best classification results with the highest accuracy.