HCFMRP COVID-19 & LID (v1)

Dataset designed to characterize lung interstitial diseases (LID) and COVID-19 on chest X-ray. For the composition of the dataset, frontal chest X-ray of patients from the Ribeirão Preto Medical School of University of São Paulo (Brazil) were classified into three groups: healthy (382 images), with LID (308 images), and with COVID-19 (189 images).

All images were analyzed by a thoracic radiologist and COVID-19 diagnosis was confirmed by RT-PCR.

For each of the groups, three types of files are available, all anonymized: the frontal view of the chest X-ray in DICOM format; the original images converted to PNG format; and the same images in 3-channel PNG format (RGB).

The following directory and file structure is presented:

    1-Normal: for healthy cases

        normal-dcm-anonymized: anonymized DICOM files

        normal-png: image files converted to PNG

        normal-png-RGB: image files converted to 3-channel PNG

    2-LID: for LID cases

        lid-dcm-anonymized: anonymized DICOM files

        lid-png: image files converted to PNG

        lid-png-RGB: image files converted to 3-channel PNG

    3-COVID: for COVID-19 cases

        covid-dcm-anonymized: anonymized DICOM files

        covid-png: image files converted to PNG

        covid-png-RGB: image files converted to 3-channel PNG

For more information, contact us: https://mainlab.fmrp.usp.br/