While heuristically recapitulating the face recognition using PIL and LIBSVM i discovered that the recognition probabilities are much higher if the image is first converted into YIQ color space. With PIL it is;
matrix = (0.299,0.587,0.114,0,0.596,-0.274,-0.322,0,0.212,-0.523,0.311,0)
Image = Image.convert("RGB",matrix)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
While heuristically recapitulating the face recognition using PIL and LIBSVM i discovered that the recognition probabilities are much higher if the image is first converted into YIQ color space. With PIL it is;
matrix = (0.299,0.587,0.114,0,0.596,-0.274,-0.322,0,0.212,-0.523,0.311,0)
Image = Image.convert("RGB",matrix)
Thanks for the tip. I will look into it.