Sign in to follow this  
jimgardener

normalisation doubts

Recommended Posts

hi i was trying to learn the topic of face recognition with eigenfaces using the code at (http://darnok.org/programming/face-recognition/EigenFaceComputation.java) in the code he is normalising a double[][] as follows(line num:202) for ( image = 0; image < nrfaces; image++) { temp = max(eigVector[image]); // Our max for ( pix = 0; pix < eigVector[0].length; pix++) // Normalize eigVector[image][pix] = Math.abs( eigVector[image][pix] / temp); } my understanding is that normalisation should make a vector length equal to 1 (and so is done as x/|x| ),but here the array is divided by max() of each row.. why is this?can anyone help? jim

Share this post


Link to post
Share on other sites
Vector normalization involves creating a unit vector. More generally, it's the process of making something more normal (as defined appropriately for the task).

In this case, you're boosting the dynamic range (increasing contrast, stretching the histogram). I've not done much image processing, but presumably higher-contrast details are easier to detect.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this