A Graph Cuts Image Segmentation Method for Quantifying Barrier Permeation in Bone Tissue

Abstract

Bio-imaging techniques have recently gotten a lot of attention since they have enabled in-vivo imaging. They are expected to contribute to drug discovery, understanding of disease mechanisms etc. However, data retrieved by bioimaging techniques have been increasing in volume, and it is not anymore feasible to analyze it manually. Therefore automatic extraction of characteristic of a huge amount of data have become important. Moreover, quantitative analysis methods are required for statistical reliability. In this paper we introduce a method for the analysis of a sequence of bone tissue images taken by a two-photon microscope to quantify blood permeability of bone marrow. This method segments the input image sequence to blood vessel, bone marrow and bone regions by graph cuts which we extended according to the images. Permeability is quantified by the intensity of the segmentation result. We also confirm that our method shows that quantification tendency is similar to ground truth data made by an expert.