Graph cut-based segmentation

Jens Petersen, Ipek Oguz, Marleen de Bruijne


This chapter describes how to use graph cut methods for medical image segmentation. Graph cut methods are designed to solve problems that can be modeled using Markov random fields. A brief introduction to graph theory, flow networks, and Markov Random Fields are therefore given. The chapter shows how a range of segmentation tasks can be formulated as such energy minimization problems and demonstrates how they can be solved with graph cuts. Specific examples of how to segment coronary arteries in computed tomography angiography images and the multilayered surfaces of airways in computed tomography images are given.

TitelMedical Image Analysis
RedaktørerAlejandro F. Frangi, Jerry L. Prince, Milan Sonka
Antal sider27
Publikationsdato1 jan. 2024
ISBN (Trykt)9780128136584
ISBN (Elektronisk)9780128136577
StatusUdgivet - 1 jan. 2024


Dyk ned i forskningsemnerne om 'Graph cut-based segmentation'. Sammen danner de et unikt fingeraftryk.