[Retracted] An Enhanced Priori Knowledge GAN for CT Images Generation of Early Lung Nodules with Small-Size Labelled Samples

Table 1

The lung nodule detection results using faster R-CNN on synthesized data generated by SLS-PriGAN and DCGAN.

Train data	Method	Precision	Recall
G_dataset1	SLS-PriGAN	0.91	0.81	0.85
G_dataset2	DCGAN	0.88	0.78	0.82
R_dataset	—	0.92	0.90	0.89

G_dataset1 and G_dataset2 are synthesized datasets generated by SLS-PriGAN and DCGAN, respectively. R_dataset is the real CT image data without any modification or synthesized data, which, in this case, can be considered as ground truth to benchmark with two other methods.