0

I have already extracted image features of 3 classes using SURF (openCV) and stacked all descriptors together into a numpy array.

My next steps would be:

  1. clustering using K-means
  2. build feature histogram for each training case based on clusters
  3. train a multi class linear SVM to separate cases based on feature histograms

My question is: at what point would it be most appropriate to scale my data?

Should this be done before clustering, histograms or SVM?

I am using sklearn's standard scaler, as well as linearSVC.

pepe
  • 9,799
  • 25
  • 110
  • 188
  • 1
    This might help http://cs231n.github.io/neural-networks-2/ – Ash May 17 '16 at 01:45
  • thanks for the great resource -- but are those operations done on the extracted features or on the pixels values? – pepe May 17 '16 at 12:58
  • Related http://stackoverflow.com/questions/14383034/opencv-combining-surf-with-neural-network – Ash May 18 '16 at 04:47

0 Answers0