Model-Based Segmentation, Reconstruction and Analysis of the Vertebral Body from Spinal CT - Spinal Imaging and Image Analysis - page 382

Fig. 9 The VB separation:
The green volume shows the
4 four seeds. The red lines
correspond to the number of
voxels whose gray level are
bigger than 200 HU. The
black lines correspond to the
threshold written in Eq. ( 6 )

The proposed method produced about 85 % successful separation results, if
excellent and good grades are considered. Hence, 15 % separation results were
considered as fair, bad, or fail.

2.4 Segmentation Using Sign Distance-based Shape Model,
Gaussian-based Intensity Model, and Symmetric Gibbs
Potential-based Spatial Interaction Model

In the following subsections, we describe three proposed methods developed for the
VB segmentation problem. We,

first, describe each method, then show its experi-

mental results.
The overall segmentation framework is shown in Fig. 10 . The proposed
framework steps are described in Algorithm 1 as follows:

Algorithm 1:

A(Training)): 80 training VB shapes are used to obtain the new proba-
bilistic shape model. In this step, manually segmented VB shape which
are obtained from 20 different patients and different regions (such as
cervical, thoracic, and lumbar spine bone sections) are used.