Skip to content

Latest commit

 

History

History
31 lines (24 loc) · 1.76 KB

segmentation_nuclei.md

File metadata and controls

31 lines (24 loc) · 1.76 KB

Nuclei segmentation

Nuclei segmentation in 3D data is challenging because of background intensity, uneven intensity in Z-dimension, noise and simply the amoung of pixels which need to be processed. Real-time experience while configuring a workflow for nuclei segmentation can be achieved when utilizing classical methods such as filtering, thresholding and watershed techniques. It is recommended to utilize modern GDDR6-based GPU hardware for 3D segmentation.

How to do 3D cell nuclei segmentation

Open your data set. Start the CLIJx-Assistant and follow such a workflows:

  • Your dataset
    • CLIJx-Assistant Starting point
      • [Optional: Noise removal and Background subtraction]
        • Threshold DoG
          • Parametric Watershed
            • Connected Components Labeling
              • Maximum Z projection

After assembling your workflow, put these operations next to each other, change the parameters.

<iframe src="https://clij.github.io/clijx-assistant/images/incubator_segmentation_3d_nuclei.mp4" width="540" height="540"></iframe> [Download video](https://clij.github.io/clijx-assistant/images/incubator_segmentation_3d_nuclei.mp4) [Image data source: Daniela Vorkel, Myers lab, CSBD / MPI CBG]

There are many ways for detecting nuclei and extending their size, e.g. to study neighborhood relationships.

<iframe src="https://clij.github.io/clijx-assistant/images/clijxa_teaser1_fast.mp4" width="800" height="640"></iframe> [Download video](https://clij.github.io/clijx-assistant/images/clijxa_teaser1.mp4) [Image data source: Daniela Vorkel, Myers lab, CSBD / MPI CBG]

Back to CLIJx-Assistant

Imprint