Table 1.

Examples of datasets with medical images.

Name	Description	Reference
BRATS	The Multimodal Brain Tumor Segmentation Benchmark (BRATS) is an annual challenge that aims to compare different algorithms for brain tumor segmentation. The dataset, which has received several enhancements over the years, consists of preoperative multimodal MRI scans of glioblastoma and lower-grade glioma with ground truth labels and survival data for participants to segment and predict the tumor.	[118]
KiTS	The Kidney Tumor Segmentation Benchmark (KiTS) is a dataset used to evaluate and compare algorithms for kidney tumor segmentation. The dataset consists of CT scans of the kidneys and kidney tumors, with 300 scans in total. The data and segmentations are provided by various clinical sites around the world.	[119]
LiTS	The Liver Tumor Segmentation Benchmark (LiTS) is a dataset used to evaluate and compare liver tumor segmentation algorithms. The dataset consists of CT scans of the liver and liver tumors, with 130 scans in the training set and 70 scans in the test set. The data and segmentations are provided by various clinical sites around the world.	[94]
MURA	The Musculoskeletal Radiographs (MURA) dataset is a large dataset of musculoskeletal radiographs containing 40,561 images from 14,863 studies. Each study is manually labeled by radiologists as either normal or abnormal.	[120]
MedPix	A free online medical image database with over 59,000 indexed and curated images from over 12,000 patients.	[121]
NIH Chest X-rays	A large dataset of chest X-ray images containing over 112,000 images from more than 30,000 unique patients. The images are labeled with 14 common disease labels.	[122]