Segmentation of head bones in 3-D CT images from an example

Abstract

We propose a new generic framework for segmentation of 3D digital data, based on knowledge contained in a segmentation example of similar data. The integration of prior knowledge is made by registering the image to segment on the segmentation example. Since the registration step relies on binary segmented data, segmentation and registration are performed jointly in a coarse-to-fine way using a multiscale parametric representation of a threshold map and of a deformation field. The threshold map is required by the segmentation procedure which is also devoted to recover topological details. The benefit of such an approach is illustrated in the context of head bone segmentation in 3-D computed tomography (CT) images.

Binary Image Registration Based on Geometric Moments: Application to the Registraion of 3d Segmented Ct Head Images

“…We compare the proposed scheme to the symmetrized version of the ICP registration method. 3 Results are presented in Table 1. Each row corresponds to the registration result for a GMI of a given degree and radius.…”

“…14 An alternative solution is to symmetrize the ICP criterion in order to match the voxels belonging to the object component and the voxels belonging to the background component. 3 The distance is then composed of two terms: the first one corresponds to the distance computed with the ICP algorithm whereas the second term is similar to the ICP distance, except that the computation is based on background voxels. In the result section, the proposed algorithm will be compared with this modified version of the ICP algorithm.…”

“…However, in the case of CT scan skull images, gray level information is limited, the contrast is low and the intensity values are heavily distorted. 3 Consequently, gray level information is not representative of the anatomical properties and is not reliable to distinguish image features. In such cases, assuming extraction of the image objects is available, binary image registration is a common alternative.…”

Topology Preserving Warping of Binary Images: Application to Atlas-Based Skull Segmentation

“…Another perspective is to consider the proposed framework for devising strategies whose behaviour may evolve during the deformation process, for example by performing in parallel segmentation and registration, as proposed in [9].…”

Interactive 3D brain vessel segmentation from an example

“…However, a compromise between automatic and interactive segmentation could be proposed, thanks to the notion of example-based segmentation, which has been considered in several application fields [2,3] before being applied to medical imaging [4]. The use of segmentation examples can lead to an automatic presegmentation of the regions where are located the vessels, and in particular to determine a subnetwork of interest into the vascular network, e.g., for vein/artery separation in images where both structures appear.…”