Multiple hypothesis tracking based extraction of airway trees from CT data: using statistical ranking of template-matched hypotheses

TY - CONF

T1 - Multiple hypothesis tracking based extraction of airway trees from CT data

T2 - Medical Imaging Summer School 2016

AU - Raghavendra, Selvan

AU - Petersen, Jens

AU - de Bruijne, Marleen

PY - 2016

Y1 - 2016

N2 - Segmentation of airway trees from CT scans of lungs has important clinical applications, in relation to the diagnosis of chronic obstructive pulmonary disease (COPD). Here we present a method based on multiple hypothesis tracking (MHT) and template matching, originally devised for vessel segmentation, to extract airway trees. Idealized tubular templates are constructed and ranked using scores assigned based on the image data. Several such regularly spaced hypotheses are used in constructing a hypothesis tree, which is then traversed to obtain improvedsegmentation results.

AB - Segmentation of airway trees from CT scans of lungs has important clinical applications, in relation to the diagnosis of chronic obstructive pulmonary disease (COPD). Here we present a method based on multiple hypothesis tracking (MHT) and template matching, originally devised for vessel segmentation, to extract airway trees. Idealized tubular templates are constructed and ranked using scores assigned based on the image data. Several such regularly spaced hypotheses are used in constructing a hypothesis tree, which is then traversed to obtain improvedsegmentation results.

M3 - Poster

Y2 - 31 July 2016 through 6 August 2016

ER -