*This dataset supports online use.Click here to jump.

Lung Phantom Dataset is a lung lesion dataset. It is based on the chest model produced by the US Food and Drug Administration. It produces 12 lesion shadows of different sizes (effective diameter 10 and 20mm), shapes (spherical, elliptical, lobed and burr-shaped) and densities (-630, -10 and +100 HU), and is scanned and recorded using a scanner from Columbia University-Medical Center. The CT scanning parameters are 120 kVp, 100 mAs, collimation 64*0.625, spacing 1.375mm, and 1.25mm thick slices are used as lung core reconstructed images.

The Lung Phantom Dataset was released by the Cancer Imaging Archive (TCIA) in 2015. The related paper is “Data From Lung_Phantom:The Cancer Imaging Archive”.