1、皮肤分割和面部检测数据集 FSD
（1）数据集名称：Face and Skin Detection (FSD) Database （2D图像）
（2）数据集简介：
The Face and Skin Detection (FSD) Database is created to support research on skin segmentation and face detection. It contains 4,000 colour images that are diverse in terms of the background scenes, lighting conditions, and face and skin types. The lighting conditions include indoor lighting and outdoor lighting; the skin types include whitish, brownish, yellowish, and darkish skins. To provide the ground-truth, all images are meticulously segmented for face regions and skin regions. The skin segmented images consist of all exposed skin regions such as facial skin, neck, arms, and hands.
（2）地址：
Son Lam Phung （需要邮件申请才能下载）；
https://documents.uow.edu.au/~phung/docs/face_skin_database.pdf
（3）标注信息：面部分割的语义分割标注；皮肤切割的语义分割标注；
（4）数据概况
（1）数据集名称：lapa-dataset（2D图像）
（2）数据集地址：
下载地址（谷歌云盘）：https://drive.google.com/file/d/1XOBoRGSraP50_pS1YPB8_i8Wmw_5L-NG/view?usp=sharingx
人脸数据集Lapa（训练集part2-1）-深度学习文档类资源-CSDN下载人脸数据集Lapa（训练集part2-2）-深度学习文档类资源-CSDN下载
（3）标注信息：头部各部位的语义分割标注
（4）数据概况：22000张（2.2G）
（1）数据集名称： All-Age-Faces (AAF) Dataset （2D图像）
（2）数据集地址：
下载地址：
Download link (baidu): All-Age-Faces Dataset.zip_免费高速下载|百度网盘-分享无限制 Download link (dropbox): https://www.dropbox.com/s/a0lj1ddd54ns8qy/All-Age-Faces%20Dataset.zip?dl=0
（3）标注信息：年龄、性别
The annotation files in folder
image setshave the following format:"%05dA%02d %d\n", person_id, age, gender,where for gender, 0 stands for female and 1 stands for male.
（4）数据概况：
图片像素低，大概都是在500*500以下，没法获取精细的面部特征。
The All-Age-Faces (AAF) Dataset contains 13'322 face images (mostly Asian) distributed across all ages (from 2 to 80), including 7381 females and 5941 males.
Each image contains a different individual, and is given a unique name (
%05dA%02d.jpg), illustrating the individual’s serial number and specific age. Individuals from serial number 00000 to 07380 are all female, from 07381 to 13321 are male.
（1）数据集名称： Celebrity-Face-Recognition-Dataset（2D图像）
（2）数据集地址：
主页：https://github.com/prateekmehta59/Celebrity-Face-Recognition-Dataset
下载地址（谷歌云盘）：https://drive.google.com/drive/folders/0B5G8pYUQMNZnLTBVaENWUWdzR0E?usp=sharing
（3）标注信息：人名分类标签+1个未知类别
Dataset of around 800k images consisting of 1100 Famous Celebrities and an Unknown class to classify unknown faces.
（4）数据概况：
Dataset of around 800k images consisting of 1100 Famous Celebrities and an Unknown class to classify unknown faces.
All the images have been scraped from Google and contains no duplicate images. Each Celebrity class(folder) consists approximately 700-800 images and the Unknown class consists of 100k images.
Total Size of the Dataset - 172 GB. Total number of zipped files - 12
（1）数据集名称：Flickr-Faces-HQ (FFHQ)
（2）数据集简介：期初是因为对抗网络任务而收集的
（2）数据集地址：
下载地址（谷歌云盘）：详见主页的链接
下载地址（CSDN）：高清人脸数据集Flickr-Faces-HQ_FFHQ\\00000-2-深度学习文档类资源-CSDN下载
Flickr-Faces-HQ_FFHQ/00000-1-深度学习文档类资源-CSDN下载
（3）标注信息：无
（4）数据概况：下载中
The dataset consists of 70,000 high-quality PNG images at 1024×1024 resolution and contains considerable variation in terms of age, ethnicity and image background. It also has good coverage of accessories such as eyeglasses, sunglasses, hats, etc.
（1）数据集名称：A Face Detection Benchmark
（2）数据集简介：
（2）数据集地址：
下载地址：
test：文件分享
dev：
train：
（3）标注信息：
（4）数据概况：
FDDB dataset contains the annotations for 5, 171 faces in a set of 2, 845 images.
AFW dataset is built using Flickr images. It has 205 images with 473 labeled faces. For each face, annotations include a rectangular bounding box, 6 landmarks and the pose angles.
PASCAL FACE consists of 851 images and 1, 341 annotated faces
IJB-A [14] is proposed for face detection and face recognition. IJB-A contains 24, 327 images and 49, 759 faces.
The IJBA dataset has large quantity of labeled data; however, occlusion and pose are not annotated.
1、全称
2、数据概况：我能下载到的全部都是黑白的数据，而且图像都很小，大概300×300左右
12、MALF
MALF is the first face detection dataset that supports fine-gained evaluation. MALF [34] consists of 5, 250 images and 11, 931 faces.
The IJBA dataset has large quantity of labeled data; however, occlusion and pose are not annotated.
