Sovereign 8s.is mirror of standardgalactic/BubbleSegmentation
  • Jupyter Notebook 99.7%
  • Python 0.3%
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2023-07-18 08:44:06 +01:00
__pycache__ Finished model 2023-07-18 00:19:36 +01:00
.DS_Store Initial setup 2023-07-03 17:12:04 +01:00
.gitattributes Initial commit 2023-07-03 17:07:21 +01:00
.gitignore Initial setup 2023-07-03 17:12:04 +01:00
bubble_processing.ipynb More fixes 2023-07-18 08:25:13 +01:00
bubble_test_notebook.ipynb Finished model 2023-07-18 00:19:36 +01:00
bubbles.py More fixes 2023-07-18 08:25:13 +01:00
README.md Create README.md 2023-07-18 08:44:06 +01:00
threshold_testing.py Finished model 2023-07-18 00:19:36 +01:00
utils_visualisation.py More fixes 2023-07-18 08:25:13 +01:00

BubbleSegmentation

BubbleSegmentation is software that identifies bubbles in images, using a scale-equivariant convolutional kernel. This results in a bubble heatmap for each radius, which is thresholded to pick out a 'constellation' of points at each bubble centre. The threshold level varies as a function of radius and bubble volume fraction, which was derived empirically from sample data.

Repository guide

  1. bubbles.py contains all the main functions
  2. utils_visualisation.py has utility and visualisaiton functions
  3. bubble_processing.ipynb is the guided notebook for using the software
  4. bubble_test_notebook.ipynb is the working test notebook
  5. bubble_data is the folder where bubble_processing.ipynb will save the computed bubble data
  6. threshold_testing.py has the empirical measurements used to create the thresholding model

How to use BubbleSegmentation

  1. Clone the repository and create a new folder called 'data' inside the main one. Put all your images in subfolders in there. I.e. the paths of your images should look like 'data/05.17_0.37/230517_AR_1632.jpg'.
  2. Install openCV if you haven't already. The other dependencies are standard.
  3. Open bubble_processing.ipynb and follow the instructions, working through each experiment folder at a time.