# Need some help with dicom CT Images

**URL:** <https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47>\
**Category:** Uncategorized\
**Created:** [June 9, 2021, 6:55pm UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47 "2021-06-09T18:55:16Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![hasan](https://yyz2.discourse-cdn.com/free1/user_avatar/shapeworks.discourse.group/hasan/32/19_2.png) [@hasan](https://shapeworks.discourse.group/u/hasan)\
**Post date:** [June 9, 2021, 6:55pm UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/1 "2021-06-09T18:55:16Z")

</div>

![image](https://global.discourse-cdn.com/free1/uploads/shapeworks/original/1X/3512e42d49ba9e0ec7ce16958d97e041da0f02ca.jpeg)

Hello good people,

I am trying to use ShapeWorks for my thesis purpose, but I can’t proceed with it ☹ basically, I tried the ellipsoid dataset to run codes on Jupiter notebooks. But when I was trying to use my dataset, which is CT scan images in Dicom format, I could not convert it to vtk format. I am attaching a screenshot here.

Also, my primary aim is to generate mesh from dicoms. Would anyone help me by showing the right direction to do that? Some tutorials or guidelines, maybe?

Also I tried them in ShapeworkStudio 6.0

 ![image](https://global.discourse-cdn.com/free1/uploads/shapeworks/original/1X/054213d1e055cf90cccd5bb4c18ddffce677a93e.jpeg)

It’s not showing anything. What could possibly go wrong?

Thank you so much.

---

<div class="post-metadata">

**Author:** ![shireen](https://yyz2.discourse-cdn.com/free1/user_avatar/shapeworks.discourse.group/shireen/32/9_2.png) [@shireen](https://shapeworks.discourse.group/u/shireen)\
**Post date:** [June 14, 2021, 12:15pm UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/2 "2021-06-14T12:15:45Z")

</div>

Hi Hassan,

Are your Dicom images original CT scans or segmented anatomy? ShapeWorks needs the geometry of the object of interest you would like to model and analyze. It can not work directly from image scans.

Regards  
Shireen

---

<div class="post-metadata">

**Author:** ![hasan](https://yyz2.discourse-cdn.com/free1/user_avatar/shapeworks.discourse.group/hasan/32/19_2.png) [@hasan](https://shapeworks.discourse.group/u/hasan)\
**Post date:** [June 15, 2021, 7:00am UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/3 "2021-06-15T07:00:34Z")

</div>

yeah, they are segmented.

anyway, please take a look.

 ![image](https://global.discourse-cdn.com/free1/uploads/shapeworks/original/1X/caf1570fbf49792ee36493f29a9d2e67be67baf0.png)

---

<div class="post-metadata">

**Author:** ![shireen](https://yyz2.discourse-cdn.com/free1/user_avatar/shapeworks.discourse.group/shireen/32/9_2.png) [@shireen](https://shapeworks.discourse.group/u/shireen)\
**Post date:** [June 15, 2021, 1:54pm UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/4 "2021-06-15T13:54:42Z")

</div>

Could you try to convert dicoms to single-file formats such as nrrd or mha? We haven’t tested the shapeworks read-image command with dicoms yet but it might worth a try at your end. You can use the shapeworks read-image and write-image commands to convert between image formats. Another option is to use ITKSnap (free software) to export dicoms to nrrds.

Meanwhile, can you share a sample image with us? @amorris could help with troubleshooting this on our side.

---

<div class="post-metadata">

**Author:** ![amorris](https://avatars.discourse-cdn.com/v4/letter/a/73ab20/32.png) [@amorris](https://shapeworks.discourse.group/u/amorris)\
**Post date:** [June 15, 2021, 8:20pm UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/5 "2021-06-15T20:20:06Z")

</div>

The `shapeworks` read-image command can read DICOM images, just pass the name of the folder as the filename. You can convert directly to a mesh with a command such as:

shapeworks read-image --name dcm\_file\_folder image-to-mesh write-mesh --name file.vtk

---

<div class="post-metadata">

**Author:** ![hasan](https://yyz2.discourse-cdn.com/free1/user_avatar/shapeworks.discourse.group/hasan/32/19_2.png) [@hasan](https://shapeworks.discourse.group/u/hasan)\
**Post date:** [June 17, 2021, 2:01am UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/6 "2021-06-17T02:01:50Z")

</div>

Here is a sample dataset. Please follow this link. I’ve uploaded it in my personal google drive and you can access it with this link-

[drive link](https://drive.google.com/drive/folders/10iYL6a3fcjayI-nmKUd5buy_K22ABdxj?usp=sharing)

Please check it.

Thanks!

---

<div class="post-metadata">

**Author:** ![shireen](https://yyz2.discourse-cdn.com/free1/user_avatar/shapeworks.discourse.group/shireen/32/9_2.png) [@shireen](https://shapeworks.discourse.group/u/shireen)\
**Post date:** [June 24, 2021, 11:34pm UTC](https://shapeworks.discourse.group/t/need-some-help-with-dicom-ct-images/47/7 "2021-06-24T23:34:40Z")

</div>

These files are CT images. See below.

 ![image](https://global.discourse-cdn.com/free1/uploads/shapeworks/original/1X/9152ec1350c782c34fe8592a4e7849fe1a407173.jpeg)

It is not about the file format, dicom vs nrrd. Shape modeling with ShapeWorks (or any other shape modeling tool) starts with the segmentation of anatomy of interest. Segmentations can either be binary volumes or surface meshes and ShapeWorks supports both. See below for illustration.

 ![](https://global.discourse-cdn.com/free1/uploads/shapeworks/original/1X/a2acd1b16f69f324d0d10e71fe7cec7dbd4776fb.png)
