I finished my internship in the company this week , I left QingDao intelligent manufacturing data Co., Ltd . The main tasks of this week are code handover and deploying my previous face recognition code to personal computer , And build the corresponding environment and test , Completed .
This week, the public data set of face recognition was also collected , give the result as follows ：

1. casia-facev5：

CASIA-FaceV5 Asian face datasets have 500 people 、 everyone 5 A picture , common 2500 A picture , The size of the picture is 640*480. The format of the data is more in line with the direction I want to do now , But there are too few photos for everyone, only 5 Zhang . link ：https://pan.baidu.com/s/1ia6LdbXv7LOAr9XirntUsg , Extraction code ：BXW2.

2. CMUVASC & PIE Face dataset

This data set contains data from 68 Individual 40000 A picture , It also includes everyone's 13 Kind of posture conditions ,43 Light conditions and 4 A picture with an expression , Website is http://vasc.ri.cmu.edu/idb/html/face/index.html, But it failed . The resources I collected online are inconsistent with the descriptions , Although every picture is made with a name annotation, But the number of photos of everyone is very small , There are no 40000 photos , And the photos are black and white , Some photos may be multi-target , Can not use ：

3.FERET

200 people , Everyone 7 Zhang , Classified , grayscale ,80x80 Pixels . The first 1 This is a standard unchanged image , The first 2、5 This is an image of a large range of attitude changes , The first 3、4 This is an image of small amplitude attitude change . Use it as a training set （ Comparison Library ） Just right , But no test set .
link ：https://pan.baidu.com/s/1o9Y8v9rcdHn9763Z3HnVqA Extraction code ：v7pt

4.CAS-PEAL

CAS-PEAL Institute of computing technology, Chinese Academy of Sciences 2003 Completed in includes 1040 Volunteer workers 99450 A database of face pictures . The database covers posture 、 expression 、 decorate 、 light 、 background 、 Changes in characteristics such as distance and time , It is a data set that meets my expectations very well , Divide the training set and the test set ,
Download link ：https://pan.baidu.com/s/1KZtOk-sfNkxnNQfNsTgTrQ Extraction code ：jbp6

5.CASIA-WebFace

CASIA-WebFace The author wrote a crawler software , Take, climb, take IMDb And save it , The number of photos of everyone is different

6. AFLW

annotation There are a lot of dirty data , Although everyone may have multiple photos , But it cannot be distinguished by label files .

7.GENKI4K

Smiley face dataset , File corruption , The photo cannot be opened

8.The Japanese Female Facial Expression (JAFFE) Database

213 Zhang image ,10 personal . Everyone in a group , Each group contains 7 Kind of expression , Each expression probably has 3,4 A sample picture . It can be regarded as annotation

link ：https://pan.baidu.com/s/1hrICsVq password ：klve

9.FDDB

Face Detection Data Set and Benchmark, It is an open database maintained by the computer department of the University of Massachusetts , Provide a standard face detection and evaluation platform for researchers from all over the world . It is one of the most authoritative face detection and evaluation platforms in the world , contain 2845 A picture , share 5171 Personal face as test set . Download address ：https://data.vision.ee.ethz.ch/cvl/rrothe/imdb-wiki/

10.JAFFE

The database is created by 10 Japanese women made various expressions according to instructions in the experimental environment . The whole database has 213 Zhang image ,10 personal , It's all women , Everyone makes 7 Kind of expression .

11. Person identification in TV series

Face photos are from two TV series 《 Vampire Hunter Buffy 》 and 《 The big bang theory 》, The label is mat File format , The picture is videvens Format file , It's troublesome to use .

https://cvhci.anthropomatik.kit.edu/~baeuml/downloads/CVPR2013_PersonID_data_v1.1.tar.bz2

11. Personal data sets

The comparison library is as follows

The test set is the face image obtained from the surveillance camera , as follows ：

summary

at present The Japanese Female Facial Expression (JAFFE) Database、CASIA-WebFace、CAS-PEAL、 and Personal data sets It can be used , One of the most expected is CAS-PEAL