Journal of Big Data

Table 2 Performance of EGAN when batch size is varied for all evaluated datasets

From: Unsupervised feature learning-based encoder and adversarial networks

Dataset	Classifier	Batch size
Dataset	Classifier	8	16	32	64
Tea clones	AlexNet	92.92	94.02	92.47	91.15
Tea clones	DenseNet	94.24	94.46	92.25	94.02
PV corn	AlexNet	91.26	93.25	91.66	92.85
PV corn	DenseNet	96.03	95.63	96.42	96.42
PV potato	AlexNet	94.85	97.05	94.11	97.05
PV potato	DenseNet	98.52	98.52	96.32	97.79
PV apple	AlexNet	88.26	93.87	91.83	92.85
PV apple	DenseNet	96.93	95.91	97.95	95.40
MNIST	AlexNet	97.86	97.70	97.63	97.03
MNIST	DenseNet	98.43	98.80	98.63	98.43

The best performance for each dataset and classifier is printed in bold

Back to article page