Automated segmentation and analysis of tree-like structures from 3D medical images are important for many medical applications, such as those dealing with blood vasculature or lung airways. However, there is an absence of large databases of expert segmentations and analyses of such 3D medical images, which impedes the validation and training of proposed image analysis algorithms. In this work, we simulate volumetric images of vascular trees and generate the corresponding ground truth segmentations, bifurcation locations, branch properties, and tree hierarchy. The tree generation is performed by iteratively growing a vascular structure based on a user-defined (possibly spatially varying) oxygen demand map. We describe the details of the algorithm and provide a variety of example results.

No abstract

Abstract. The problem of scarcity of ground-truth expert delineations of medical image data is a serious one that impedes the training and validation of medical image analysis techniques. We develop an algorithm for the automatic generation of large databases of annotated images from a single reference dataset. We provide a web-based interface through which the users can upload a reference data set (an image and its corresponding segmentation and landmark points), provide custom setting of parameters, and, following server-side computations, generate and download an arbitrary number of novel ground-truth data, including segmentations, displacement vector fields, intensity nonuniformity maps, and point correspondences. To produce realistic simulated data, we use variational (statistically-based) and vibrational (physically-based) spatial deformations, nonlinear radiometric warps mimicking imaging nonhomogeneity, and additive random noise with different underlying distributions. We outline the algorithmic details, present sample results, and provide the web address to readers for immediate evaluation and usage.

In previous work, we presented an algorithm to synthesize volumetric images of vascular trees and generate the corresponding ground truth segmentations, bifurcation locations, branch properties, and tree hierarchy. In this work, we provide the software needed to simulate these volumes. Our software expects a number of physical parameters and oxygen demand maps to produce 3D volumetric images of vasculature, as well as information about the bifurcation locations, tree hierarchy and branch radii in a GXL file. We foresee our software useful for large scale evaluation studies of medical image segmentation and analysis software.