Premium
Image segmentation by graph cut for radiation images of small animal blood vessels
Author(s) -
Kim Taewan,
Ahn Chibum,
Lee Onseok
Publication year - 2018
Publication title -
microscopy research and technique
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.536
H-Index - 118
eISSN - 1097-0029
pISSN - 1059-910X
DOI - 10.1002/jemt.23154
Subject(s) - cut , segmentation , artificial intelligence , graph , image segmentation , computer vision , computer science , pattern recognition (psychology) , anatomy , biology , mathematics , combinatorics
Synchrotron radiation (SR) based X‐ray imaging is an attractive method for analyzing biomedical structure. However, despite its many advantages, there are few gold standards in image processing methods, especially in segmentation. Image segmentation is an essential step in medical imaging for image analysis and three‐dimensional reconstruction. Although there are many algorithms for image segmentation, a decisive method does not exist in SR X‐ray imagery, because of a lack of data. This study focused on finding a suitable algorithm for image segmentation in high‐resolution medical imaging. In this study, we used following four algorithms to segment blood vessel of mouse; interactive graph cuts algorithm, which segments an image using fast min‐cut/max‐flow algorithm to solve global solution, binary partition tree algorithm, which uses an interactive method creating tree nodes to segment an image by using splitting and merging an image, seeded region growing algorithm, which performs segmentation by connecting similar pixel value, simple interactive object extraction, which generates color signature for segmentation based on color model of an image.