Deep Transfer Learning for Biology Cross-Domain Image Classification

Table 4

Accuracy (%) results of training from scratch.

Model	Flowers17	Flowers102	Plant Seedlings	PlanktonSet 1.0	QUT Fish
AlexNet	83.82	47.02	96.21	75.09	25.84
VGG-16	84.12	44.35	97.08	76.19	25.21
GoogLeNet-v3	87.94	55.44	97.48	77.12	32.92
ResNet-18	90.29	55.60	98.02	76.36	34.96
ResNet-34	89.71	57.16	97.68	76.73	36.51
ResNet-50	83.53	52.77	97.58	77.23	30.55
ResNet-101	83.24	50.76	97.47	77.33	30.06
ResNet-152	81.18	51.24	97.05	77.40	30.64

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