takaria0 / 3d_nuclei_segmentation

Spheroid 3D nucleus segmentation library written in Python

3D Spheroid nucleus segmentation software written in Python. You can easily run and get the volume, label, centroid of the nucleus in the spheroid and many more.

How to use

What you need is

• Python Environment (Python 3.9.0 or higher)
• 3D image of a spheroid (usually .tif)

Install

1. Install Python 3.9.0 or higher
2. Clone this repository
3. In the directory, run the following command to install required dependencies

pip3 install -r requirements.txt

Set settings.json

see settings_example.json

1. You need to set at least these properties to run the segmentation

• "img_path": ".../spheroid.tif", Input 3D image, make sure it is single-channel 3D nuclei image (.tif).
• "output_dir": ".../output", Output directory to save the segmentation results.

2. After you set the values, rename settings_example.json to settings.json.

Run

In the directory, run

python3 main.py

See the Result

Generated files are like these.

You can see the results in your output_dir. It contains .pkl file for later your analysis. The .pkl file can be loaded using Python pickle library. It contains properties that you set in the settings.json for each nucleus segmented in the image. Also, the settings.json is saved to check what parameters you used in the analysis. .npy file is a segmented 3D image in numpy format, each nucleus region is annotated as different integers. The .csv file is not needed if you use .pkl file, as it is more handy for Python. But in case you don't use Python for post processing, you can use .csv.

  "properties": [
    "area",
    "centroid",
    "axis_major_length",
    "axis_minor_length",
    "image",
    "inertia_tensor_eigvals",
    "moments_central"
  ],

For example, by using pandas.read_pickle(), you can get this segmentation results.

Debug

In settings.json, you can set debug: true and you can visualize the segmentation process step by step like below using napari library. For more info, please have a look at https://napari.org/stable/.

Troubleshooting

not working when debug is true

run

brew install PyQt