Eigenfaces

As we have broken down temperature data in the previous section, we can use a similar approach to face detection by introducing a concept called eigenfaces. We will be using AT&T Laboratories Cambridge's $($previously known as Olivetti Research Laboratory$)$ "Database of Faces" in this section. This is a set of grayscaled face images, each normalized to the same resolution and each face taking up the same relative amount of area in each image.

import eigenfaces

$($face$\_$shape, faces$)=$ eigenfaces.load$\_$face$\_$dataset$()$

num$\_$faces $=$ len$($faces$)$

print$($f$\text{'}$Loaded $\{$num$\_$faces$\}$ faces.$\text{'})$

Loaded $400$ faces.

Note that each image has the shape defined in face$\_$shape