Skip to main content

Table 1 Common data set of medical imaging report generation

From: A survey on automatic generation of medical imaging reports based on deep learning

Data set

Description

Image

Report

Link

IU X-Ray

Chest X-ray images of lung diseases

7470

3955

http://openi.nlm.nih.gov/

ChestX-ray14

14 kinds of lung diseases

112,120

–

https://nihcc.app.box.com/v/ChestXray-NIHCC

CheXpert

Chest radiographs of 65,240 patients with lung diseases

224,316

–

https://stanfordmlgroup.github.io/competitions/chexpert/

MIMIC-CXR

227,835 radiographic studies in DICOM format

377,110

227,835

https://physionet.org/content/mimic-cxr/2.0.0/

CX-CHR

Chest X-ray images with Chinese reports of 35,609 patients

45,598

–

–

PadChest

Chest X-ray data set obtained from 67,000 patients

160,000

109,931

https://bimcv.cipf.es/bimcv-projects/padchest/

PEIR Gross

Radiology teaching images

4,000

4000

https://peir.path.uab.edu/library/index.php?/category/106

DDSM

Normal, benign, and malignant mammography studies

2620

–

http://marathon.csee.usf.edu/Mammography/Database.html