Journal of Big Data

Table 2 Model performance of different transfer learning architectures in this study

From: Efficient pollen grain classification using pre-trained Convolutional Neural Networks: a comprehensive study

	Accuracy (%)	Precision (%)	Recall (%)	F1-score (%)
AlexNet	92.87	93.50	93.10	92.40
VGG-16	95.10	95.30	95.50	95.60
MobileNet-V2	94.78	96.04	95.68	95.74
ResNet-18	94.65	95.78	94.67	95.05
ResNet-34	95.32	94.78	94.57	94.47
ResNet-50	95.37	96.58	96.50	96.50
ResNet-110	96.84	96.86	96.28	96.43
ResNeSt-50	96.54	96.79	96.80	96.76
ResNeSt-101	97.24	97.89	97.13	96.86
SEResNeXt	97.05	97.66	97.31	97.01
ViT	95.95	95.71	95.46	95.54

The italic emphasis shows the most promising modeling performance

Back to article page