From the actual paper:

[…]

We generated 16 images of female genitalia by manipulating four individual photographs of the human female vulva, such that we had four colour conditions for four different base images. We used colours within the normal range expected for human genitalia, a gradation of increasing redness starting from a pale pink. This was done to prevent any aversion in our participants (described below) to ‘unnatural’ images [26]: for instance, some NHP studies of genital colour preference have contrasted red with anatomically atypical block colours (e.g. green, purple, orange) [25].

Explicit images of anatomically normal, un-retouched, non-pornographic, similarly-orientated female genitals were surprisingly difficult to obtain, and the number of images used in the experiment reflects this difficulty. We obtained the photographs from a female genital image comparison website (www.vulvavelvet.org), the purpose of which is to inform and educate the public about natural variation in human female genitalia, and to make women feel comfortable with their bodies. Women, over the age of 18, anonymously donate images of their genitals to this website, placing them in the public domain. […]

Don’t ya just love anthropology?