Adaptive multi-level conditional random fields for detection and segmentation of small enhanced pathology in medical images

Zahra Karimaghaloo, Douglas L. Arnold, Tal Arbel

Abstract: Highlights•Introducing a probabilistic Adaptive Multi-level Conditional Random Fields (AMCRF) to address the problem of small enhanced pathology segmentation.•Incorporating higher order cliques to better model the variable interactions.•Exploring the effect of multiple higher order textural patterns in order to detect structures of interest.•Investigating the effect of several different parameter learning and inference algorithms for the proposed graphical model.•Testing the proposed model on large multi-center clinical trials from Relapsing-Remitting MS patients where results show 90% sensitivity with 16% false detection rate.