This post is to describe my work presented in the article Patch-based Image Classification through Conditional Random Field Model ( link ).
The paper presents a system to perform image classification combining a technique for colour-based segmentation ( cartooning , performed via anisotropic diffusion) with the Conditional Random Field inference framework to model patches relationships. The idea is to obtain a coarse segmentation thus easing the learning stage of classes that are discriminable via their colour.
This post is just a test to see the capabilities of this blog. I may use it to store and possibly share some consideration about the papers I publish in citeUlike.