Applying Artificial Neural Networks for Face Recognition

Algorithm 1

Input shape 𝑋{(𝑥_𝑖,𝑦_𝑖)}

Output new shape 𝑋^′{(𝑥^′_𝑖,𝑦^′_𝑖)}

 For each point (𝑥_𝑖,𝑦_𝑖) of shape X

  For each sub-window 𝑠𝑤^′ centered at point (𝑥^′,𝑦^′)

  of the window centered at the feature point (𝑥_𝑖,𝑦_𝑖).

   (i) Computing MLP (𝑠𝑤^′,𝑊). If the return value is (+1) then point (𝑥^′,𝑦^′) is at the edge.

   (ii) Selecting the nearest point (𝑥^′,𝑦^′) to the point

     (𝑥_𝑖,𝑦_𝑖) as the new feature point (𝑥^′_𝑖,𝑦^′_𝑖).