Disponible en français bientôt...

It is believed there are two underlying factors that cause us to find a face beautiful. They are (a) the golden ratio, along with symmetry & (b) how closely the sample face is to a theoretical average. This program measures both to a varying degree.

The Golden Ratio

 

The measurements compared are as follows:

A = Horizontal width of the nose 58

B = Horizontal width of the mouth 90

C = Horizontal distance from corner of mouth to edge of face 61

M = Horizontal width of eye 59

N = Horizontal distance between eyes 61 (/2)

 

X = Vertical distance between lowest point of eye & chin 192

Y = Vertical distance between lowest point of nose & chin 119

The ancient Greeks calculated that, for a theoretically perfect face, distance B would be 1.618 times the distance A. Distance A would also be the same as C, 1/(2N) and M (squared?!). For this reason, 1.618 was called Phi, the golden ratio. Width of head (temple to temple) = phi cubed.

 

The Average Face

There is another theory that the closer a face is to 'the average face' the more we find it beautiful. It is quite obvious that we might find faces with very large noses or very small eyes as strange and unattractive, so the converse would appear to be a reasonable assumption.

This program takes all the co-ordinates from all the faces and calculates average positions. From these, it can compare a given set of co-ordinates & make a judgement on how close to or far away the given face lies from the average face. A close match being considered more beautiful than a distant mismatch.

 

Recognition

It is believed that we subconsciously recognise faces from the features that lie within a triangular area in the centre of the face. The corners of this triangle are given by the pupils of the eye and the middle of the mouth. As the nose lies entirely within this triangle, it is believed the nose & the distances from it to the triangle's corners are the most important. Conversely, features outside of this triangle such as eyebrows, eyelashes, the chin, the jawline etc are less important in recognition.

The software uses this fact and assigns importance 'weights' to these features, over and above the user-definable importance preferences.