bob.med.tb.data.padchest_RS

Padchest TB dataset for computer-aided diagnosis

A large chest x-ray image dataset with multi-label annotated reports. This dataset includes more than 160,000 images from 67,000 patients that were interpreted and reported by radiologists at Hospital San Juan (Spain) from 2009 to 2017, covering six different position views and additional information on image acquisition and patient demography.

We keep only “PA” images here and only the “Tuberculosis” subset with an equivalent number of “normal” images.

• Reference: [PADCHEST-2019]

• Original resolution: variable, original size

• Labels: [PADCHEST-2019]

• Split reference: 64%/16%/20%

• Protocol default:

  • Training samples: 160

  • Validation samples: 40

  • Test samples: 50

• Protocol ìdiap`: * Images path adapted to Idiap infrastructure

• Labels: DensenetRS predictions

bob.med.tb.data.padchest_RS.dataset = <bob.med.tb.data.dataset.JSONDataset object>

Padchest dataset object