Model-Based Segmentation, Reconstruction and Analysis of the Vertebral Body from Spinal CT - Spinal Imaging and Image Analysis - page 383

Fig. 10 The general segmentation framework. (Prior to this framework, it is required to obtain the
shape model). In the first phase, the spinal cord is extracted, processes, and ribs roughly using the
Matched filter. Also, the data size is reduced to minimize the execution time. In the second phase,
the VBs are separated with two choices: (i) manual, (ii) automatic. In the third phase, a new shape
based ICM method is proposed to segment the VBs

B) Spinal cord extraction (Pre-processing): The Matched filter is used

to detect the spinal cord. This step roughly extracts the ribs and spinal
processes. Also, the data size is reduced to 120x120xZ from 512x512xZ
where Z is the number of slices. This step reduces the execution time of
the segmentation process. The output of this phase is used in the following
steps.
C) Separation of VBs (Pre-processing): Two choices are given for the
user(s)-i) manual selection of disk to obtain each VB in a datasets, ii) fully
automatic VB separation using the histogram based information. It should
be noted all steps of the framework are fully automated except this step.
D) Segmentation: Three models are used to segment VBs: The intensity,
spatial interaction, and shape models. The following ' While ' loop is

processed for the segmentation.

While j \ N slices do (N slices : the number of slices.)