Validation Metrics for Non-rigid Registration of Medical Images Containing Vessel Trees

Validation of non-rigid registration methods is still a chal- lenging task. Different evaluation criteria were published, yet no widely accepted gold standard exists. The aim of this paper is to provide quan- titative evaluation metrics suited for clinical 3D images containing vessel trees, such as liver or brain data. We present a method to identify corre- sponding points on different vessel trees by extracting consistent graph minors interactively. Four different metrics based on these corresponden- cies are proposed.