Improved Classification of White Blood Cells with the Generative Adversarial Network and Deep Convolutional Neural Network

Table 7

10-fold cross-validation accuracy of the DNN models initialized using random weights.

Model	GAN_aug1 (100 inst./class) (%)	GAN_aug2 (150 inst./class) (%)	GAN_aug3 (200 inst./class) (%)
VGG-16	91.9	92.6	93.4
VGG-19	92.6	93.1	93.5
ResNet-18	92.7	94.0	94.6
ResNet-50	93.8	94.5	94.9
DenseNet-121	95.0	95.6	95.7
DenseNet-169	95.3	95.4	95.8