Dataset Introduction
The constructed RSNA 2019+ dataset, based on the popular RSNA 2019 dataset, contains 421,070 CT slices with bleeding from 11,149 patients, and 175,125 bounding boxes annotated for five ICH subtypes, which is the largest ICH CT dataset with category and localization annotations for each CT slice in the world.
The CQ500+ dataset is constructed on the CQ500 dataset, containing 2,586 CT slices with bleeding from 491 patients, and 4,707 bounding boxes annotated for five ICH subtypes.At present, only demo data can be downloaded, and all data can be freely downloaded after the paper is published.
Demo Dataset
| Filename |
Size |
Download |
| Demo images.zip |
528.51MB |
Download |
| Demo labels.zip |
4.01MB |
Download |
RSNA 2019+ Training
| Filename |
Size |
Download |
| RSNA 2019 plus training images.zip |
2.77GB |
Download |
| RSNA 2019 plus training images.z01 |
4.00GB |
Download |
| RSNA 2019 plus training images.z02 |
4.00GB |
Download |
| RSNA 2019 plus training images.z03 |
4.00GB |
Download |
| RSNA 2019 plus training labels.zip |
127.69MB |
Download |
RSNA 2019+ Test
| Filename |
Size |
Download |
| RSNA 2019 plus test images.zip |
1.12GB |
Download |
| RSNA 2019 plus test images.z01 |
4.00GB |
Download |
| RSNA 2019 plus test labels.zip |
46.38MB |
Download |
CQ500+
| Filename |
Size |
Download |
| CQ500+ images.zip |
658.39MB |
Download |
| CQ500+ labels.zip |
8.73MB |
Download |