Ask Your Question

Classification algorithm in face recognition methods

asked 2016-09-07 22:29:55 -0600

Kelvin gravatar image

I'm using some algorithms (eigenfaces, fisherfaces and LBPH) to perform face recognition using the opencv contrib modules.

Can someone explain to me what internal algorithms these methods use to perform the classification? (for example K-NN, etc)

Note: if it uses K-NN, what is the default value for K?


edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2016-09-08 00:49:52 -0600

berak gravatar image

all 3 algorithms are using a simple (1) Nearest Neighbour classification, like this:

for each image in trainset:
    if distance between train & test img is the smallest so far:
          keep best id & distance
return best id & distance
edit flag offensive delete link more


Great. Thanks a lot @berak.

Kelvin gravatar imageKelvin ( 2016-09-09 19:38:30 -0600 )edit

One more thing I forgot to ask. What distance measurement is used? It uses euclidean distance?

Kelvin gravatar imageKelvin ( 2016-10-04 10:35:25 -0600 )edit

yes, exactly.

berak gravatar imageberak ( 2016-10-04 10:39:51 -0600 )edit

Ok, thanks again @berak.

Kelvin gravatar imageKelvin ( 2016-10-04 10:49:34 -0600 )edit

Question Tools

1 follower


Asked: 2016-09-07 22:29:55 -0600

Seen: 247 times

Last updated: Sep 08 '16