# Totalsegmentator --- practise data

**URL:** <https://discourse.slicer.org/t/totalsegmentator-practise-data/29428>\
**Category:** Development\
**Tags:** segmentation\
**Created:** [May 12, 2023, 12:52pm UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428 "2023-05-12T12:52:10Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![whu](https://avatars.discourse-cdn.com/v4/letter/w/7ab992/32.png) [@whu](https://discourse.slicer.org/u/whu)\
**Post date:** [May 12, 2023, 12:52pm UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428/1 "2023-05-12T12:52:10Z")

</div>

After installing the Totalsegmentator extensions successfully.  
I found little problem in the visualization.  
Firstly, when using the example CTA data, when “Show 3D”, the result is:  
 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/2/2/22d4e33d00396f253ee98bf71d28fb177ff35a77.jpeg)  
and the local detail is like this,seemed not so smooth as it should be.  
 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/5/2/529887822ae86d8455f1ac63730161e54a16d067.jpeg)  
Second, when using the whole heart ct data (the famous segmentation dataset:[Multi-Modality Whole Heart Segmentation 2017](https://zmiclab.github.io/zxh/0/mmwhs/data.html)),one of the dataset:  
the process is good,but the visualization result is the same:  
 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/7/a/7ada4380db8d14819c790402c0b4e9e5dcebb1fc.jpeg)  
If anyone want to practice it , I would like to share one of the dataset.

---

<div class="post-metadata">

**Author:** ![rbumm](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/rbumm/32/9404_2.png) [@rbumm](https://discourse.slicer.org/u/rbumm)\
**Post date:** [May 12, 2023, 4:30pm UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428/2 "2023-05-12T16:30:17Z")

</div>

You are probably using the FAST option automatically, in which the dataset gets downsampled?  
What are you system specs (GPU)?

---

<div class="post-metadata">

**Author:** ![whu](https://avatars.discourse-cdn.com/v4/letter/w/7ab992/32.png) [@whu](https://discourse.slicer.org/u/whu)\
**Post date:** [May 14, 2023, 12:46pm UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428/3 "2023-05-14T12:46:44Z")

</div>

thanks.  
Win10 64bits，i7 12cores, NVIDIA Quadro P620（2G graphics memory）。  
So the totalsegmentator force to use fast mode, it seemed that the graphics memory should be more then 7G.

---

<div class="post-metadata">

**Author:** ![rbumm](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/rbumm/32/9404_2.png) [@rbumm](https://discourse.slicer.org/u/rbumm)\
**Post date:** [May 14, 2023, 3:05pm UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428/4 "2023-05-14T15:05:47Z")

</div>

You could install Pytorch in cpu mode, I guess you would get better segmentation results but would have a longer waiting time.

---

<div class="post-metadata">

**Author:** ![whu](https://avatars.discourse-cdn.com/v4/letter/w/7ab992/32.png) [@whu](https://discourse.slicer.org/u/whu)\
**Post date:** [May 15, 2023, 10:45am UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428/5 "2023-05-15T10:45:11Z")

</div>

the console version in Aconda virtual enviroment,Totalsegment seemed can not support “Show 3D”, just the slice by slice UNet segmentation. 😄

---

<div class="post-metadata">

**Author:** ![whu](https://avatars.discourse-cdn.com/v4/letter/w/7ab992/32.png) [@whu](https://discourse.slicer.org/u/whu)\
**Post date:** [May 15, 2023, 11:23am UTC](https://discourse.slicer.org/t/totalsegmentator-practise-data/29428/6 "2023-05-15T11:23:27Z")

</div>

the heart\_myocardium segmentation result is like this  
 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/1/3/137422a5f510e0dcacad34fda97898885a6e48af.png)  
~