denote the intensity non uniformity corrected image [8] with intensities in the normalized intensity space , where
are the coordinates in the continuous image domain
Throughout the whole section, we furthermore assume that the tubular structure of interest is roughly oriented parallel to the
axis. Figuring out the orientation should be straightforward for most clinical applications, as the subject’s orientation with respect to the image can be determined from the image’s meta data for most clinical imaging modalities.
Original max-flow formulation. A general formulation for the continuous max-flow problem with spatial flow
, source flow
, sink flow
, and corresponding flow capacities
,
is stated by Yuan et al. [6] as

subject to the flow capacity constraints

and the flow conservation constraint








(1)

(2)

(3)
2.1 Cross-Sectional Similarity Prior
Following our goal to impose a cross-sectional similarity prior on the segmentation, we split the spatial flow
into an in-slice component
and a through-slice component
with respect to slices that lie perpendicular to the
axis (see Fig. 1a). The resulting continuous max-flow problem can then be written as follows:

subject to the new flow capacity constraints

and the new flow conservation constraint

where
denotes the divergence of
perpendicular to the
axis and
denotes the derivative of
along the
axis.






(4)

(5)

(6)







Fig. 1
Method overview. a Proposed flow configuration: the spatial flow is split into an in-slice component
, perpendicular to the axis along which the tubular structure is oriented, and a through-slice component
, parallel to the axis. b Sample sagittal slice of one of the images used for evaluation. c Segmentation result. d Surface reconstruction with cutting planes for volume measurement


The flow formulation now possesses the desired property of having the spatial flow capacity
of [6] represented by two separate terms, namely the in-slice flow capacity
and the through-slice flow capacity
. The latter capacity,
, represents the cross-sectional similarity prior that allows for precise control over the through-slice flow behavior: For example, we may choose an edge-based cost function for
that drives the segmentation towards edges in
, while setting
to enforce constant similarity throughout all slices. Or we may calculate
as a slice wise cost-function that, for each slice, adjusts the similarity prior to the in-slice noise level (reinforcing the similarity prior if the noise level is high and relaxing it if the noise level is low). Other combinations are possible, of course: note that both
and
may be formulated pointwise.










Dual formulation. Introducing the Lagrange multiplier
and following the steps in [6], the max-flow problem can be reformulated as the equivalent primal-dual model

subject to the capacity constraints (5). The equivalent dual model representing a relaxed min-cut problem then becomes
![$$\begin{aligned} \min _{u\in [0,1]}E(u)\mathbin {:=}\int \limits _{\Omega }\left\{ (1-u)C_{s}+uC_{t}+\alpha |{{\mathrm{\nabla _{{12}}}}}u|+\beta |u^{\prime }|\right\} {\mathrm{{d}}}{x}. \end{aligned}$$](/wp-content/uploads/2016/10/A331518_1_En_10_Chapter_Equ8.gif)
Here,
denotes the in-slice gradient and
denotes the through-slice derivative of
with respect to the
axis, similar to the definitions of
and
above. It can be shown that each level set function
given by
4).




(7)
![$$\begin{aligned} \min _{u\in [0,1]}E(u)\mathbin {:=}\int \limits _{\Omega }\left\{ (1-u)C_{s}+uC_{t}+\alpha |{{\mathrm{\nabla _{{12}}}}}u|+\beta |u^{\prime }|\right\} {\mathrm{{d}}}{x}. \end{aligned}$$](/wp-content/uploads/2016/10/A331518_1_En_10_Chapter_Equ8.gif)
(8)






![$$u^{\ell }(x),\,\ell \in (0,1]$$](/wp-content/uploads/2016/10/A331518_1_En_10_Chapter_IEq41.gif)

Fig. 2
Features used in segmentation. a Image intensities. b Vesselness response. c Csfness response
2.2 Tubularity Features
As our goal is to segment tubular structures in the image, it appears natural to include tubularity features in the flow capacity calculations. A well-known tubularity feature is Frangi’s measure of vesselness [9] (see Fig. 2b),
, where, for each scale
in the predefined set of scales
, the vesselness
of bright tubular structures on dark background is

with
denoting the ordered eigenvalues (
) of the pointwise Hessian matrices that result from convolving the input image
with Gaussian derivatives of standard deviation
. We define
as half of the maximum Hessian norm at the current scale as suggested by Frangi [9].





(10)





In our experiments on segmenting the spinal cord, we decided to include another feature that specifically describes the background that immediately surrounds the target structure. The spinal cord is embedded in cerebrospinal fluid (CSF), which appears dark in the used MR sequences. As the CSF also appears largely elongated, but exhibits both tube-like and plate-like properties, we adapt Frangi’s vesselness feature to a csfness feature
(see Fig. 2c) that discriminates between blob-like structures and non-blobs. We do so by replacing the eigenvalue ratio terms of
with an equivalent term composed of
and
, as it is the latter ratio that discriminates both vessels and plates from blobs in Hessian eigenvalue analysis [9]. Consequently, we define
for dark non-blobs on bright background in the scales
with

Combining the features. Let
,
be the vesselness and csfness feature spaces, let
and
be two combined feature spaces, let
,
be two new image functions that map to the combined feature spaces, and let
,
be the coordinates in the combined feature spaces.







(11)
![$$\mathcal {V}=[0,1]\ni v^{*}$$](/wp-content/uploads/2016/10/A331518_1_En_10_Chapter_IEq57.gif)
![$$\mathcal {W}=[0,1]\ni w^{*}$$](/wp-content/uploads/2016/10/A331518_1_En_10_Chapter_IEq58.gif)







Stay updated, free articles. Join our Telegram channel

Full access? Get Clinical Tree


