[Insight-users] Extract centerlines from a 3D segmented binary image

Dan Mueller dan.muel at gmail.com
Mon Apr 9 17:13:26 EDT 2012


Hi superZZ,

Good to hear the approach is suitable for your needs.

Cheers, Dan

On 10 April 2012 03:01, jxdw_zlf <jxdw_zlf at yahoo.com.cn> wrote:
> Dear Dan,
>
> Thank you! It works fine now. Really appreciate for your help :)
>
> Best Regards,
>
> superZZ
>
>
> ------------------ Original ------------------
> From:  "Dan Mueller"<dan.muel at gmail.com>;
> Date:  Mon, Apr 9, 2012 08:41 PM
> To:  "zlf"<jxdw_zlf at yahoo.com.cn>;
> Cc:  "insight-users"<insight-users at itk.org>;
> Subject:  Re: [Insight-users] Extract centerlines from a 3D segmented binary
> image
>
> Hi superZZ,
>
> To reiterate my first email: this method is a little tricky due to the
> need for a _good_ speed function. A good speed function has high
> values (~1.0) near the centerline and low values (0.0) away from the
> centerline.
>
> A distance map can be used generate such a speed function, but the
> output of the distance map is _NOT_ the speed function: the distance
> map requires further processing!
>
> Here are the steps I used to quickly generate a speed function
> (commands are in SharpImage syntax, a custom prototyping environment
> of mine, but you should be able to work out what is happening):
>
> SignedMaurerDistance
> SmoothingRecursiveGaussian Sigma=1.0
> Threshold Lower=-10 Upper=100 OutsideValue=-10
> RescaleIntensityToF OutputMinimum=0.0 OutputMaximum=1.0
> UnaryPixelMath Operation="Power" Exponent=3
>
> The results look acceptable to me (see attached).
>
> HTH
>
> Cheers, Dan
>
> On 9 April 2012 16:26, zlf <jxdw_zlf at yahoo.com.cn> wrote:
>> Dear Dan,
>>
>> I process airway.mhd with SignedMaurerDistanceMapImageFilter. Pass
>> SignedMaurerDistanceMapImageFilter's result as speed image. The result is
>> even worse.
>>
>>
>> http://itk-insight-users.2283740.n2.nabble.com/file/n7449026/centerline_SignedMaurerDistanceMapImageFilter.jpg
>>
>> Regards,
>>
>> superZZ


More information about the Insight-users mailing list