AICONSlab/MIRACL

Blank segmentation output

Open

#186 opened on Aug 23, 2025

View on GitHub
 (3 comments) (0 reactions) (3 assignees)Python (11 forks)github user discovery
help wanted

Repository metrics

Stars
 (49 stars)
PR merge metrics
 (PR metrics pending)

Description

I tried to segment my own lsm data. for starter I tried a small block but the generated segmentation are all blank.

here's the command I run in the apptainer.

miracl seg ace -sai /data/200316-1-1_N12_N58204NLSAM_NeuN-ls_M4D/chunk_0_start_1767_4325_941_end_2192_3858_1159.tif -sao /data/outputs/ -sam unet -sar 1.8 1.8 4.0

In the input folder, I have slice_z001.tif to slice_z218.tif, 218 slices in total. each slice has shape (425, 467).

I got seg_final folder, and generated_patches with.

out_patch_0_0.tiff
patch_0_0.tiff
percentage_brain_patch.json

the percentage_brain_patch.json is

{
    "patch_0_0.tiff": 7.688754051923752
}

The out_patch_0_0.tiff has size 512,512,512 and completely blank.

May I ask is there any requirements for the input data? what might cause the issue of having blank segmention output?

Contributor guide

Research direction
Check the segmentation pipeline for patch generation and prediction steps. Ensure input dimensions match model requirements. Inspect the patch creation code and verify that the model output is correctly written. Look for issues with scaling or empty patches.
Tech stack
python
Domain
datamachine learning
Issue type
Bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
Python
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
70