Eigenfaces - What are the 400 pictures used for?

asked 2016-08-19 13:07:25 -0600

Hello, I've been looking into Eigenfaces and I'm trying to understand what the 400 training images are used for in the process of detecting your face. I thought that you just have a set of images of yourself and when the picture is taken it checks against those images. What do these 400 images have to do with my face?

edit retag flag offensive close merge delete

Comments

1

I think that you need 400 negative pictures in order to be able to separate your face to the other "not your face" in the features space. You also have to retain only the disciminative features as you want to reduce the dimentionnality of the feature space, so you will have to choose which features are the best to tell that this face is your face or not your face.

More information with this tutorial or with more experienced people.

Eduardo gravatar imageEduardo ( 2016-08-19 14:26:44 -0600 )edit

the purpose of the facerecognizer classes is to find the closest of N persons in a database. (classification)

it sounds a bit, like you're not trying to solve that problem. can you be more specific about what you're trying to achieve ?

berak gravatar imageberak ( 2016-08-20 00:33:48 -0600 )edit

I'm just trying to figure out what the 400 pictures are used for because I need to be able to explain that when I'll be doing a presentation on this algorithm.

FreakySlax gravatar imageFreakySlax ( 2016-08-20 19:48:48 -0600 )edit
1

in that case, please RTFM

berak gravatar imageberak ( 2016-08-20 20:33:16 -0600 )edit