25/11/2016 - 10:00
Fast centerline detection and segmentation algorithm for anatomical tree structures
Kristı́na Lidayová
Salle de Réunion 4ème étage Blaise Pascal
Type d'Evenements: 

Kristı́na Lidayová is currently pursuing the Ph.D. at Uppsala University, Sweden. She received her BS degree in computer science and MS degree in mathematics from Comenius University, Bratislava, Slovakia in 2009 and 2012, respectively. Her current research interests include medical image analysis and segmentation.

Summary of the presentation :

Segmentation of tubular tree structures is a specific segmentation challenge that can be facilitated and accelerated by first identifying the centerlines of the tree structure. At this seminar, I will present a fully automatic high-speed skeleton extraction algorithm combined with an algorithm for a complete anatomical tree segmentation. I will explain how the algorithm works and demonstrate it on a Computed Tomography Angiography (CTA) images of the lower limbs. In addition, I will mention how the algorithm is modified to segment diseased peripheral arteries and airway trees from CT images of piglets with Acute Respiratory Distress Syndrome (ARDS).