There are lists which address not available or outdated faces datasets. Below is the list of datasets available for academic / non-commercial usage, if not stated otherwise.
- PubFig
http://www.cs.columbia.edu/CAVE/databases/pubfig/download/
dev/test stats:
60/140 individuals
16366/42k images
list of urls - Labeled Faces in the Wild
http://vis-www.cs.umass.edu/lfw/
13233 images
5749 people
1680 people with two or more images
gzipped - Mining Wild Faces in Wikipedia Biographies
http://www.info.polymtl.ca/~mdhas/LWF/
8.5k images
1.5k people
gzipped - Fasescrub
http://vintage.winklerbros.net/facescrub.html
100 000 images
530 people
list of urls - Adience
http://www.openu.ac.il/home/hassner/Adience/data.html
26 580 images
2284 people
gzipped
Tips
If you encounter problem with downloading images from provided list, you can do as follows:
$ cat urls_list_with_data.txt | awk ‘{ print $5 }’ > urls_list.txt
this command takes five column from file (the one with url) and saves to separate file facescrub_actresses_urls_list.txt, so you can download it with wget:
$ wget -i urls_list.txt
where -i options is used to get urls sequentially or:
$ cat urls_list.txt | parallel -j 20 –gnu “wget {}”
to spread task of downloading files among 20 threads.