# MONAI Auto3DSeg – inconsistent performance for vertebral body segmentation

**URL:** <https://discourse.slicer.org/t/monai-auto3dseg-inconsistent-performance-for-vertebral-body-segmentation/46081>\
**Category:** Support\
**Tags:** segmentation, python, monai-auto3dseg\
**Created:** [February 6, 2026, 9:23pm UTC](https://discourse.slicer.org/t/monai-auto3dseg-inconsistent-performance-for-vertebral-body-segmentation/46081 "2026-02-06T21:23:41Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![jchao](https://avatars.discourse-cdn.com/v4/letter/j/bc8723/32.png) [@jchao](https://discourse.slicer.org/u/jchao)\
**Post date:** [February 6, 2026, 9:23pm UTC](https://discourse.slicer.org/t/monai-auto3dseg-inconsistent-performance-for-vertebral-body-segmentation/46081/1 "2026-02-06T21:23:41Z")

</div>

Hi everyone,

I’m currently working with Prof. Ron Alkalay on a spine CT segmentation project using **MONAI Auto3DSeg** and would appreciate any advice from the community.

My task is to segment **only vertebral bodies**. I reformulated it as a **binary segmentation task (VB vs. non-VB)**, which initially improved the validation Dice to \>70%.

However, when I ran a second round of training, the performance dropped noticeably and did not recover, even though the training process itself ran normally.

Between the two runs, I changed the following parameters:

- **Resample resolution**  
From: (0.3125, 0.3125, 0.5)  
To: (0.3125, 0.3125, 1.0)

- **ROI size**  
From: (128, 128, 64)  
To: (128, 128, 96)

Other than these changes, the setup and data split were kept the same.

In both experiments, I used **Auto3DSeg’s Quick training mode**.

#### Dataset details

- `imagesTr`: ~60 GB, 257 CT volumes (`.nii.gz`)

- `labelsTr_bin`: ~385 MB, 257 binary segmentation masks (`.nii.gz`)

Some additional details:

- Tool: MONAI Auto3DSeg - segresnet\_0

- Task: Vertebral body vs. background (binary)

- Modality: CT

- Tried adjusting: ROI size, AMP, batch/auto-scaling

I’m trying to understand whether this performance drop is likely related to the coarser through-plane resolution, the larger ROI depth, Quick mode limitations, or some interaction with Auto3DSeg’s preprocessing and model selection.

If anyone has experience using Auto3DSeg for similar anatomical segmentation tasks, I would really appreciate any guidance on what to sanity-check first, or best practices for tuning these parameters.

I’m happy to share configs or logs if helpful.

Thank you very much in advance!

---

<div class="post-metadata">

**Author:** ![pieper](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/pieper/32/8_2.png) [@pieper](https://discourse.slicer.org/u/pieper)\
**Post date:** [February 8, 2026, 4:31pm UTC](https://discourse.slicer.org/t/monai-auto3dseg-inconsistent-performance-for-vertebral-body-segmentation/46081/2 "2026-02-08T16:31:36Z")

</div>

I have tried Auto3DSeg for training but not extensively. The results were good, but not noticeably better than nnU-Net, so I’ve standardized on nnU-Net for a the last few years and have generally found that it behaves predictably as more data is added for training. I don’t use any resampling or cropping, only segmentations and native resolution CT data.

The details change for each dataset, but the basic format for me is still as demonstrated [in this nnmouse repo](https://github.com/pieper/nnmouse).

The format of the input directories is a little different between Auto3DSeg and nnU-Net, but they are close enough that you can accomplish the renaming just with symbolic links, so you don’t need to duplicate the data.

Was there a particular reason you wanted to use Auto3DSeg instead of nnU-Net?