LITERATURE SURVEY OF AUTOMATIC FACE RECOGNITION ...

CHAPTER XV
CLASSIFYING THE FACES
The process of classification of a new (unknown) face Гnew to one of the
classes (known faces) proceeds in two steps.
First, the new image is transformed into its eigenface components. The
resulting weights form the weight vector Ω T new.
The Euclidean distance between two weight vectors d ( Ωi, Ω j) provides a
measure of similarity between the corresponding images i and j. If the Euclidean
distance between Гnew and other faces exceeds ­ on average ­ some threshold
value Θ, one can assume that Гnew
is no face at all. d ( Ωi, Ω j) also allows one to
construct ”clusters” of faces such that similar faces are assigned to one cluster.
is done,
Then computation of the distance between the face and its reconstruction
ξ 2 = ||rrm ­ rs || 2
After this we have to distinguish between face and non­face images, by
applying these conditions on our calculated result. The conditions are:
1. If ξ ≥ Θ,
then the image is not a face.
2. If ξ < Θ and εi ≥ Θ,
then it’s a new face.
3. If ξ < Θ and min { εi} < Θ,
then it’s a known face.
56