Related Experiment Video
Updated: Jun 25, 2025
07:01
3D Imaging of Soft-Tissue Samples using an X-ray Specific Staining Method and Nanoscopic Computed Tomography
Published on: October 24, 2019
9.8K
Automated 3D cytoplasm segmentation in soft X-ray tomography
Ayse Erozan1,2,3, Philipp D Lösel2,3,4, Vincent Heuveline2,3
1Centre for Organismal Studies, Heidelberg University, Heidelberg, Germany.
Iscience
|May 24, 2024
Summary
Automated 3D cytoplasm segmentation using ACSeg accelerates analysis of soft X-ray tomography (SXT) data. This deep learning model significantly reduces time and labor for high-throughput cell volume and structure studies.
Area of Science:
- Cell biology
- Biophysics
- Medical imaging
Background:
- Cellular structure is crucial for understanding function, diagnostics, and therapy.
- Soft X-ray tomography (SXT) offers high-resolution, label-free imaging of cellular structures.
- Manual segmentation of SXT data is a time-consuming bottleneck.
Purpose of the Study:
- To develop an automated 3D cytoplasm segmentation model for SXT data.
- To reduce the time and labor associated with SXT image analysis.
- To enable high-throughput analysis of cell volume and cytoplasmic structure.
Main Methods:
- Developed ACSeg, an automated 3D cytoplasm segmentation model.
- Utilized semi-automated labels and a 3D U-Net architecture.
- Trained the model on 43 SXT tomograms of immune T cells, with additional data from other cell types.
Main Results:
- ACSeg rapidly achieved high-accuracy segmentation on SXT data.
- The model demonstrated successful segmentation of unseen tomograms.
- Diversifying the training set with other cell types showed potential for broader applicability.
Conclusions:
- ACSeg effectively automates 3D cytoplasm segmentation for SXT.
- The model significantly accelerates SXT data analysis, reducing manual effort.
- ACSeg facilitates high-throughput studies of cell volume and structure in various cell types.
