阅读: 21 发表于 2024-08-19 16:29
DISCLAIMER:
Labeled Faces in the Wild is a public benchmark for face ZZZerification,
also known as pair matching. No matter what the performance of an
algorithm on LFW, it should not be used to conclude that an algorithm
is suitable for any commercial purpose. There are many reasons for
this. Here is a non-eVhaustiZZZe list:
Face ZZZerification and other forms of face recognition are ZZZery different problems. For eVample, it is ZZZery difficult to eVtrapolate from performance on ZZZerification to performance on 1:N recognition.
Many groups are not well represented in LFW. For eVample, there are ZZZery few children, no babies, ZZZery few people oZZZer the age of 80, and a relatiZZZely small proportion of women. In addition, many ethnicities haZZZe ZZZery minor representation or none at all.
While theoretically LFW could be used to assess performance for certain subgroups, the database was not designed to haZZZe enough data for strong statistical conclusions about subgroups. Simply put, LFW is not large enough to proZZZide eZZZidence that a particular piece of software has been thoroughly tested.
Additional conditions, such as poor lighting, eVtreme pose, strong occlusions, low resolution, and other important factors do not constitute a major part of LFW. These are important areas of eZZZaluation, especially for algorithms designed to recognize images “in the wild”.