Comparison of different metods for atlas construction

When comparing experimental data obtained from different subjects, a standard approach is to display results on an atlas, as a common anatomical substrate. Through time, different methodologies for atlas constructions have been developed. A lot of works on this topic discuss the selection of the “best” atlas, using different criteria. In this article we discuss several methods for atlas construction and propose a novel method for least-biased atlas construction. The four different methods proposed in this article are compared based to their segmentation results. The results show no significant error reduction if the same methodology but with different atlas is implemented. Finally, we propose a methodological improvement that is not based solely on “best” atlas selection but rather on “best” segmentation selection as well. The image set used in this work consist out of 140 ultrasound images of cardiac outflow velocity profiles from both healthy volunteers and patients.