# 3D cell segmentation pipeline

**URL:** <https://forum.allencell.org/t/3d-cell-segmentation-pipeline/96>\
**Category:** Miscellaneous\
**Tags:** segmentation\
**Created:** [August 6, 2019, 2:57pm UTC](https://forum.allencell.org/t/3d-cell-segmentation-pipeline/96 "2019-08-06T14:57:19Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Joel\_Luethi](https://avatars.discourse-cdn.com/v4/letter/j/8baadc/32.png) [@Joel\_Luethi](https://forum.allencell.org/u/Joel_Luethi)\
**Post date:** [August 6, 2019, 2:57pm UTC](https://forum.allencell.org/t/3d-cell-segmentation-pipeline/96/1 "2019-08-06T14:57:19Z")

</div>

You have published this really cool [Segmenter](https://www.allencell.org/segmenter.html) for subcellular structure segmentations. I also saw that you have been using images of single cells, e.g. in the [integrated cell models](https://github.com/AllenCellModeling/torch_integrated_cell) and in the Cell Viewer. I assume those were segmented single cells? What method are you using to segment cells in 3D? The Segmenter tool seems optimized for subcellular segmentation, not cell segmentation.

Also, I’ve seen that you have worked with CellProfiler and committed code to CP for 3D support. Would you be willing to share an example segmentation pipeline that you use for 3D segmentation in CellProfiler, if you have something like this?

---

<div class="post-metadata">

**Author:** ![jianxu\_chen](https://avatars.discourse-cdn.com/v4/letter/j/85f322/32.png) [@jianxu\_chen](https://forum.allencell.org/u/jianxu_chen)\
**Post date:** [August 6, 2019, 11:00pm UTC](https://forum.allencell.org/t/3d-cell-segmentation-pipeline/96/2 "2019-08-06T23:00:14Z")

</div>

Thanks for your interest. We do have segmentation of single cells. The current results, as used in integrated cells, are obtained by the method documented in the following page:

> **[Extracting Information](https://www.allencell.org/extracting-information.html)**
>
> For each gene-edited intracellular structure, the data we acquire comprises high resolution 3D z-stacks containing one gene-edited protein. We then place these structures into a frame of reference...

This was implemented in Matlab, not part of the new segmenter.

Meanwhile, we have developed a new deep learning based method for 3D cell/nuclei segmentation using our new segmenter. We haven’t publicly presented this method yet. The public release of these new methods is on our plan. Stay tuned 🙂

For CellProfiler 3d, technically, all the classic image segmentation workflows as you have seen in our segmenter could be implemented as a CellProfiler 3D pipeline. We haven’t done so due to limited bandwidth.

Again, thanks for your interest. Let me know if you have more questions.

Jianxu