0

What exactly would the benefit of having more than one eigenface to work with when trying to do facial recognition?

Say I have an average face already prepared from my set of faces, and I know that many of the faces are different pictures of the same person. I have prepared eigenfaces of each of the images. Is there any way to combine eigenfaces of the same person in order to create a vector that could identify that person more easily?

PixelArtDragon
  • 143
  • By the way: I'm not exactly sure that this is the right site for this question- this seemed the most appropriate but if it falls outside its scope, please move it. – PixelArtDragon Mar 07 '16 at 18:54
  • Facial expressions or face angle to the camera (with a hidden variable corresponding to angle). – Alex R. Mar 07 '16 at 19:22
  • That would imply that I'm checking my input face against each eigenface separately. Is there any useful combination of multiple eigenfaces so that only one check is required? – PixelArtDragon Mar 07 '16 at 19:33
  • Then I don't really understand what you mean by "prepare eigenfaces for of the images." By definition, the eigenfaces are fixed by your face collection and the only reasonable interpretation would be to write each individual face as a linear sum of eigenfaces. So what do you then mean by combining eigenfaces of the same person? – Alex R. Mar 07 '16 at 19:46

0 Answers0