MuGeminorum/HEp2

---
license: mit
task_categories:
- image-classification
language:
- en
tags:
- biology
- medical
pretty_name: HEp-2 Cell
size_categories:
- 10K<n<100K
# dataset_info:
# splits:
# - name: train
# num_examples: 10876
# - name: validation
# num_examples: 1360
# - name: test
# num_examples: 1360
---

# Dataset card for "MuGeminorum/HEp2"
The HEp-2 (Human Epithelial type 2) dataset is a widely utilized benchmark in the field of medical image analysis, particularly for the task of antinuclear antibody (ANA) pattern classification. This dataset comprises microscopic images of HEp-2 cells stained with fluorescent dyes, showcasing diverse patterns of autoantibody binding associated with various autoimmune diseases. Researchers and practitioners leverage the HEp-2 dataset to develop and assess algorithms for automating ANA pattern recognition, thereby aiding in the diagnosis of autoimmune disorders. The intricate patterns within the dataset challenge the robustness of computational models, making it a valuable resource for advancing the understanding of autoimmune diseases and contributing to the development of cutting-edge medical image analysis techniques.

## Usage
```python
from datasets import load_dataset

data = load_dataset("MuGeminorum/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_LFS_SKIP_SMUDGE=1 git clone [e-mail]:datasets/MuGeminorum/HEp2
```

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

## Reference
[1] [Chapter III ‐ Classifying Cell Images Using Deep Learning Models](https://github.com/MuGeminorum/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>

原始信息汇总

HEp-2 Cell Dataset Summary

基本信息

• 许可证: MIT
• 任务类别: 图像分类
• 语言: 英语
• 标签: 生物学, 医学
• 数据集名称: HEp-2 Cell
• 数据集大小: 10K<n<100K

数据集描述

HEp-2 (Human Epithelial type 2) 数据集是医学图像分析领域广泛使用的基准，特别用于抗核抗体(ANA)模式分类任务。该数据集包含使用荧光染料染色的HEp-2细胞的显微图像，展示了与各种自身免疫疾病相关的抗体结合的不同模式。研究人员和实践者利用此数据集开发和评估自动化ANA模式识别算法，辅助自身免疫疾病的诊断。数据集中的复杂模式对计算模型的鲁棒性提出了挑战，使其成为推动自身免疫疾病理解和尖端医学图像分析技术发展的宝贵资源。

数据集划分

• 训练集: 10876个样本
• 验证集: 1360个样本
• 测试集: 1360个样本