Automated localisation and boundary identification of superficial femoral artery on MRI sequences

Abstract

In this paper, an automated method to localise the right superficial femoral artery (SFA) and identify its boundary on magnetic resonance imaging (MRI) sequences without contrast medium injection is proposed. Some anatomical knowledge combined with the mathematical morphology is used to distinguish SFA from other vessels. Afterwards, the directional gradient, continuity and the local contrast are applied as features to identify the artery's boundary using dynamic programming. The accuracy analysis shows that the system has average unsigned errors 3.1 ± 3.1% on five sequences compared to experts' manual tracings.

Keywords::

• superficial femoral artery
• boundary detection
• dynamic programming

Acknowledgements

This work was supported, in part, by the German Research Foundation (‘Deutsche Forschungsgemeinschaft’, DFG), under Grants SCHU 2514/1-1 and SCHU 2514/1-2 and by the National Science Council (NSC), Taiwan, under Grant NSC 100-2221-E-039-001.

Competing interests: The authors have no competing interests.

Authors' contributions: DCC and TCH have contributions on engineering aspects including: designed the software system of SFA wall detection, wrote the most part of the manuscript, and performed all experiments shown in this paper. AST and UHS have contributions on medical aspects including: designed the medical experiments, built the gold standard (manual drawings of SFA wall), and helped in manuscript writing. All authors have read and approved the final manuscript.