# Deep Vein Segmentation And 3D Reconstruction

**URL:** <https://discourse.slicer.org/t/deep-vein-segmentation-and-3d-reconstruction/14027>\
**Category:** Development\
**Tags:** segmentation\
**Created:** [October 14, 2020, 1:34am UTC](https://discourse.slicer.org/t/deep-vein-segmentation-and-3d-reconstruction/14027 "2020-10-14T01:34:03Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Sedra\_Mulki](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/sedra_mulki/32/8439_2.png) [@Sedra\_Mulki](https://discourse.slicer.org/u/Sedra_Mulki)\
**Post date:** [October 14, 2020, 1:34am UTC](https://discourse.slicer.org/t/deep-vein-segmentation-and-3d-reconstruction/14027/1 "2020-10-14T01:34:03Z")

</div>

Hello

I am working on 3D slicer to extract deep vein in legs by using MRI images.

And I used 2 way to get the result:

1- First way I apply Level set segmentation from Vascular Modeling Toolkit Model,

 ![SceneView](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/4/1/4132253f54e57b0eb1706a20033c5744889aaabf.jpeg)

2- second way I apply Region Growing Model,

 ![2020-06-04-Scene](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/7/2/72fb59fd0e1fafe258318ce155fac5d9d30792dd.jpeg)

My first question is: this model that I applied is ok for segmentation of deep vein.  
My second question is: I try to get more accuracy result, So there is model or way in 3d slicer to calculate the accuracy or (error)

Thank you,

Sidra

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [October 14, 2020, 2:02am UTC](https://discourse.slicer.org/t/deep-vein-segmentation-and-3d-reconstruction/14027/2 "2020-10-14T02:02:19Z")

</div>

There is no such thing as a perfect segmentation, only that is suitable for a specific clinical purpose. What is your goal? What information would you like to extract from the image? What are your accuracy requirements?

---

<div class="post-metadata">

**Author:** ![Sedra\_Mulki](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/sedra_mulki/32/8439_2.png) [@Sedra\_Mulki](https://discourse.slicer.org/u/Sedra_Mulki)\
**Post date:** [October 16, 2020, 7:09pm UTC](https://discourse.slicer.org/t/deep-vein-segmentation-and-3d-reconstruction/14027/3 "2020-10-16T19:09:26Z")

</div>

Hello,

Thank you for your reply.

My goal is to segment MRI images for deep vein in legs to get the boundary of vein in each image , after that reconstruct the vein by using boundaries in all images.  
I applied Level Set Segmentation from VMTK model to get the geometry of deep vein in knee.  
Also I applied Simple Region Growing Segmentation for same previous reason.  
So I need to know how I can calculate the error between real vein and vein that I extract?  
For example the diameter/volume of vein that I extract is equal to diameter/volume of real vein or how much the vein that I extract is match to real vein?  
What are the criteria to detect the error?