FFHQ-Wrinkle
该数据集包含1,000张人工标注的人脸图像及其皱纹遮罩对,以及50,000张弱标注图像,数据模态为图像,主要用于面部皱纹分割任务,支持计算机视觉和医学成像领域的研究,如皮肤分析、皱纹检测和模型训练。

基本信息
资源简介
FFHQ-Wrinkle
【数据集背景】
本数据集源站标题为《labhai/ffhq-wrinkle-dataset》。源站首次发布:2024-05-22。
【数据内容】
源站原始描述:[ICPR 2024] Official dataset repository for the paper "Facial Wrinkle Segmentation for Cosmetic Dermatology: Pretraining with Texture Map-Based Weak Supervision" ## Flickr-Faces-HQ-Wrinkle Dataset (FFHQ-Wrinkle) FFHQ-Wrinkle is an extension of the FFHQ (Flickr-Faces-HQ) dataset, specifically designed to include additional features related to facial wrinkles. This dataset aims to support research and development in facial recognition, aging simulation, and other related fields. If you use this da…
【获取与许可】
源站页面:https://github.com/labhai/ffhq-wrinkle-dataset
使用许可:NOASSERTION。
【使用注意】
· 源站最后修改于 2025-12-16,引用时请注意版本时效。
下载信息
注册下载
需要注册 Kaggle 账号并登录后下载,适合需要跟踪下载记录和使用 API 的用户。
暂未开放公开下载
无需注册即可直接获取公开样本或文档,适合快速预览和评估数据集质量。
免登录有偿下载
公开数据集受托下载与技术交付服务。
提供高速下载与技术交付服务(收技术服务费,非数据销售)
暂未开放当前数据集主要来源为 Kaggle 公开托管,完整影像包建议通过原始链接或 Kaggle API 获取。
使用方式
数据集获取
git clone https://github.com/labhai/ffhq-wrinkle-dataset.git
curl -L -o repo.zip https://github.com/labhai/ffhq-wrinkle-dataset/archive/refs/heads/main.zip
unzip repo.zip
源站 README 摘录(使用方式)
Usage
All scripts must be run from the root folder of the repository (ffhq-wrinkle-dataset).
Wrinkle label download
To download the wrinkle labels consisting of 1,000 manual wrinkle masks and 50,000 weak wrinkle masks, please run download_ffhq_wrinkle.sh. This script will download both the manual wrinkle masks and weak wrinkle masks into the specified base_folder.
bash download_ffhq_wrinkle.sh
You can set the default download folder by modifying the base_folder variable within the download_ffhq_wrinkle.sh.
Additionally, you can directly download the labels via the provided Google Drive link.
The folder structure after following the instructions is as follows:
{base_folder}/
├── manual_wrinkle_masks/
│ ├── 00001.png
│ ├── 00011.png
│ ├── ...
│ └── 21035.png
└── weak_wrinkle_masks/
├── 00000/
│ ├── 00000.png
│ ├── 00001.png
│ ├── ...
│ └── 00999.png
├── 01000/
│ ├── 01000.png
│ ├── 01001.png
│ ├── ...
│ └── 01999.png
├── ...
└── 49000/
├── 49000.png
├── 49001.png
├── ...
└── 49999.png
Face image download
To download the original FFHQ face images, please refer to the FFHQ Dataset website. We use the ‘‘images1024x1024’’ subset with IDs from 00000 to 49999 from the original FFHQ dataset. After downloading the images, place them in the designated base_folder.
The folder structure after downloading face images is as follows:
{base_folder}/
├── images1024x1024/
│ ├── 00000/
│ │ ├── 00000.png
│ │ ├── 00001.png
│ │ ├── ...
│ │ └── 00999.png
│ ├── ...
│ └── 49000/
│ ├── 49000.png
## 数据加载示例(图像类)
```python
from PIL import Image
import glob, os
files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
+ glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
+ glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)
# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path) # 要求 子目录=类别
目录组织与标注格式以源站说明和下载后实际文件为准。
数据缺失?
依托陕西、四川两大基地,我们与超过十家三甲医院建立直接合作关系,覆盖合规授权、采集治理、专业标注、数据交付的全流程,为AI医疗团队提供即拿即用的高质量临床数据。
⚡️ 需要数据支持或标注服务?立即联系我们获取专业方案。




