BOWTrainer 'blends' image features

asked 2015-11-02 14:19:14 -0600

albertJ gravatar image

I'm using the BOWTrainer class to cluster multiple segments of texture classes to create a texton dictionary and class models. For this i'm first using the .add method to add the 1d texture segments to the object then using the .cluster method to generate the cluster centres.

I'm having issues with classifier accuracy and wanted to confirm whether this class works by clustering the segments separately? I'm worried that if i'm adding textures with distinct lines,(like bricks), that these could loose their definition if they're clustered together?


edit retag flag offensive close merge delete


This question is very hard to understand, can you plz rephrase it. What do you want to cluster? What are your descriptors? What is their dimension?

Guanta gravatar imageGuanta ( 2015-11-03 11:06:40 -0600 )edit

you should not cluster images, but "features" e.g. SIFT,Gabor,HOG, where the means is more significant.

berak gravatar imageberak ( 2015-11-03 11:52:21 -0600 )edit