Datasets:

chehablab
/

NIHChestXR

patient_id stringlengths 1 5	scan_id int32 0 183	age int32 1 414	sex stringclasses 2 values	labels listlengths 1 9
1	0	58	M	[ 2 ]
1	1	58	M	[ 2, 11 ]
1	2	58	M	[ 2, 3 ]
2	0	81	M	[ 0 ]
3	0	81	F	[ 14 ]
3	1	74	F	[ 14 ]
3	2	75	F	[ 14 ]
3	3	76	F	[ 14, 4 ]
3	4	77	F	[ 14 ]
3	5	78	F	[ 14 ]
3	6	79	F	[ 14 ]
3	7	80	F	[ 14 ]
4	0	82	M	[ 5, 6 ]
5	0	69	F	[ 0 ]
5	1	69	F	[ 0 ]
5	2	69	F	[ 0 ]
5	3	69	F	[ 0 ]
5	4	70	F	[ 0 ]
5	5	70	F	[ 0 ]
5	6	70	F	[ 4 ]
5	7	70	F	[ 3, 4 ]
6	0	81	M	[ 0 ]
7	0	82	M	[ 0 ]
8	0	69	F	[ 2 ]
8	1	70	F	[ 0 ]
8	2	73	F	[ 6 ]
9	0	73	M	[ 11 ]
10	0	84	F	[ 4 ]
11	0	75	M	[ 3 ]
11	1	75	M	[ 0 ]
11	2	75	M	[ 0 ]
11	3	75	M	[ 0 ]
11	4	75	M	[ 0 ]
11	5	75	M	[ 4 ]
11	6	75	M	[ 1 ]
11	7	75	M	[ 4 ]
11	8	75	M	[ 0 ]
12	0	76	M	[ 3, 5 ]
13	0	61	M	[ 0 ]
13	1	60	M	[ 11, 8 ]
13	2	60	M	[ 11, 8 ]
13	3	60	M	[ 13 ]
13	4	60	M	[ 3, 11, 4, 8 ]
13	5	60	M	[ 11, 4, 13, 8 ]
13	6	60	M	[ 3, 4 ]
13	7	60	M	[ 4 ]
13	8	60	M	[ 0 ]
13	9	60	M	[ 13 ]
13	10	60	M	[ 3, 7, 8 ]
13	11	60	M	[ 8 ]
13	12	60	M	[ 8 ]
13	13	60	M	[ 8 ]
13	14	61	M	[ 0 ]
13	15	61	M	[ 0 ]
13	16	61	M	[ 0 ]
13	17	61	M	[ 0 ]
13	18	60	M	[ 3, 4, 8 ]
13	19	62	M	[ 0 ]
13	20	62	M	[ 8 ]
13	21	62	M	[ 3 ]
13	22	56	M	[ 4, 5 ]
13	23	57	M	[ 4, 5, 8 ]
13	24	57	M	[ 5 ]
13	25	56	M	[ 2, 4, 5, 6 ]
13	26	57	M	[ 2, 11 ]
13	27	57	M	[ 2, 3, 11, 5 ]
13	28	57	M	[ 2, 3, 11, 5 ]
13	29	57	M	[ 2, 3, 11, 5 ]
13	30	57	M	[ 1, 2, 11, 5, 8 ]
13	31	57	M	[ 11, 5 ]
13	32	57	M	[ 11, 5, 8 ]
13	33	56	M	[ 8 ]
13	34	56	M	[ 11, 8 ]
13	35	56	M	[ 1, 8 ]
13	36	56	M	[ 8 ]
13	37	56	M	[ 2, 11, 8 ]
13	38	56	M	[ 0 ]
13	39	56	M	[ 8 ]
13	40	56	M	[ 2, 11, 8 ]
13	41	56	M	[ 11 ]
13	42	56	M	[ 11 ]
13	43	56	M	[ 5, 13 ]
13	44	56	M	[ 2, 5, 13 ]
13	45	56	M	[ 2 ]
13	46	56	M	[ 4 ]
14	0	61	F	[ 0 ]
15	0	70	M	[ 0 ]
16	0	71	M	[ 0 ]
17	0	77	M	[ 3, 4, 6 ]
17	1	77	M	[ 0 ]
17	2	78	M	[ 0 ]
18	0	75	M	[ 0 ]
19	0	66	M	[ 1, 3, 13 ]
20	0	53	M	[ 13 ]
20	1	47	M	[ 12, 4 ]
20	2	48	M	[ 12, 4, 13 ]
21	0	78	M	[ 6 ]
21	1	80	M	[ 4 ]
22	0	48	M	[ 0 ]
22	1	49	M	[ 12 ]

End of preview. Expand in Data Studio

This dataset is derived from the NIH Chest X-ray Dataset (ChestX-ray14), one of the largest publicly available collections of chest radiographs, released by the NIH Clinical Center. The original collection comprises 112,120 frontal-view chest X-ray images from 30,805 unique patients, collected as part of routine clinical care. Each image was assigned one or more of 14 common thoracic disease labels, text-mined from the associated radiology reports using natural language processing (estimated label accuracy >90%). Here, we provide a Hugging Face ready version of the dataset with patient and scan-level metadata alongside each image.

📦 Dataset Structure

Each entry corresponds to a single chest X-ray image:

patient_id → Unique identifier for the patient
scan_id → Unique identifier for the individual scan/study
age → Patient age at time of scan
sex → Patient sex
labels → One or more of the 14 thoracic disease labels (or "No Finding")
image → Frontal-view chest X-ray image

🏷️ Labels / Classes

The labels field contains one or more of the following 14 disease classes, or No Finding if none are present:

Index	Class
0	No Finding
1	Atelectasis
2	Cardiomegaly
3	Effusion
4	Infiltration
5	Mass
6	Nodule
7	Pneumonia
8	Pneumothorax
9	Consolidation
10	Edema
11	Emphysema
12	Fibrosis
13	Pleural Thickening
14	Hernia

⚙️ Preprocessing

Original DICOM/PNG frontal-view X-rays were retained as provided by NIH
Patient, scan, and demographic fields were parsed into a flat tabular schema
Disease labels were kept as multi-label string fields matching the original NLP-derived annotations
Images were resized to 244×244

🚀 Usage

from datasets import load_dataset
import matplotlib.pyplot as plt

ds = load_dataset("chehablab/NIHChestXR", split="train")
sample = ds[314]
img = sample["image"]

plt.imshow(img, cmap="gray")
plt.title(f"Patient {sample['patient_id']} | Scan {sample['scan_id']} | Age {sample['age']} | Sex {sample['sex']} | Labels {sample['labels']}")
plt.axis("off")
plt.show()

📚 Citation

If you use this dataset, please acknowledge Chehab Lab and cite the original CVPR 2017 paper:

@inproceedings{Wang_2017,
  doi = {10.1109/cvpr.2017.369},
  url = {https://doi.org/10.1109/cvpr.2017.369},
  year = 2017,
  month = {jul},
  publisher = {IEEE},
  author = {Xiaosong Wang and Yifan Peng and Le Lu and Zhiyong Lu and Mohammadhadi Bagheri and Ronald M. Summers},
  title = {{ChestX}-Ray8: Hospital-Scale Chest X-Ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases},
  booktitle = {2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}
}

Original data download site: https://nihcc.app.box.com/v/ChestXray-NIHCC

📜 License

This dataset is released under the Creative Commons CC0 1.0 Universal (CC0 1.0) Public Domain Dedication. You may copy, modify, distribute, and use the data, even for commercial purposes, without asking permission.

Chehab Lab @ 2026

Downloads last month: 54

Collection including chehablab/NIHChestXR

Medical Imaging

Collection

A collection for medical imaging datasets and models • 13 items • Updated 5 days ago • 1