# C-NMC 2019: 12,528 segmented B-lymphoblast cell images, leukemic (B-ALL) versus normal, from 101 subjects (ISBI 2019, TCIA)

## Facts

- Provider: Shiv Mourya, Sandeep Kant, Pramod Kumar, Anubha Gupta, Ritu Gupta; repackaged by Chehab Lab
- License: CC-BY-3.0 (https://www.cancerimagingarchive.net/collection/c-nmc-2019/)
- Size: 3 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/chehablab/C-NMC-2019
- Access: public, readable without an account
- DOI: https://doi.org/10.7937/tcia.2019.dc64i46r
- Category: Health & Biomedical
- Source: https://doi.org/10.7937/tcia.2019.dc64i46r
- Page: https://scigantic.com/archive/47d58599-a75b-5a81-9338-49f3250e20fe
- Keywords: Health & Biomedical, Parquet, Hugging Face

## About

C-NMC 2019 (Classification of Normal versus Malignant Cells) is the dataset of the ISBI 2019 challenge on B-cell acute lymphoblastic leukemia (B-ALL), a blood cancer of white blood cells. It contains single B-lymphoblast cells segmented from stain-normalized microscopy images of blood smears and labeled by an expert oncologist, each a 450x450 RGB image on a black background, labeled as leukemic blast (ALL) or normal B-lymphoid precursor (HEM). This copy merges the three labeled training folds (10,661 cells from 73 subjects: 47 ALL, 26 normal) and the labeled preliminary test set (1,867 cells from 28 subjects), 12,528 images from 101 subjects in all; the final test set had no public labels. Columns give the original filename, partition, patient ID and class; images were converted losslessly from BMP to PNG. Leukemic and normal cells look very alike, so split by patient. One Parquet file under data/, about 500 MB (measured 2026-11-21); read with Hugging Face datasets. Use for leukemia cell classification and hematopathology models. Original data on The Cancer Imaging Archive (DOI 10.7937/tcia.2019.dc64i46r), released under CC BY 3.0. Cite Mourya et al., ALL Challenge dataset of ISBI 2019, The Cancer Imaging Archive (2019).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("chehablab/C-NMC-2019")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/train-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/47d58599-a75b-5a81-9338-49f3250e20fe
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/47d58599-a75b-5a81-9338-49f3250e20fe
- Over MCP: `get_schema_card` with id `47d58599-a75b-5a81-9338-49f3250e20fe` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
