---
license: mit
task_categories:
- image-classification
language:
- en
tags:
- biology
- medical
pretty_name: HEp-2 Cell
size_categories:
- 10K<n<100K
viewer: false
---

# Dataset card for "MuGemSt/HEp2"
The HEp-2 (Human Epithelial type 2) dataset is a widely used benchmark in the field of medical image analysis, especially for the task of antinuclear antibody (ANA) pattern classification. The dataset contains microscopic images of HEp-2 cells stained with fluorescence, demonstrating multiple patterns of autoantibody binding associated with various autoimmune diseases. The HEp-2 dataset is utilized by researchers and practitioners to develop and evaluate algorithms for automated ANA pattern recognition to aid in the diagnosis of autoimmune diseases. The intricate patterns in this dataset test the robustness of computational models, making it a valuable resource for advancing the understanding of autoimmune diseases and the development of advanced medical image analysis techniques.

## Viewer
<https://www.modelscope.cn/datasets/MuGemSt/HEp2/dataPeview>

## Usage
```python
from datasets import load_dataset

data = load_dataset("MuGemSt/HEp2")
trainset = data["train"]
validset = data["validation"]
testset = data["test"]
labels = testset.features["label"].names

for item in trainset:
    print("image: ", item["image"])
    print("label name: " + labels[item["label"]])

for item in validset:
    print("image: ", item["image"])
    print("label name: " + labels[item["label"]])

for item in testset:
    print("image: ", item["image"])
    print("label name: " + labels[item["label"]])
```

## Maintenance
```bash
git clone git@hf.co:datasets/MuGemSt/HEp2
cd HEp2
```

## Mirror
<https://www.modelscope.cn/datasets/MuGemSt/HEp2>

## Reference
[1] [Chapter III ‐ Classifying Cell Images Using Deep Learning Models](https://github.com/MuGemSt/Medical_Image_Computing/wiki/Chapter-III-%E2%80%90-Classifying-Cell-Images-Using-Deep-Learning-Models)<br>
[2] <a href="https://arxiv.org/pdf/1504.02531v1.pdf">HEp-2 Cell Image Classification with Deep Convolutional Neural Networks</a>