Note: This README is currently being revised. Some sections, particularly dependency lists, may have gotten lost and are therefor incomplete.
An adapted and extended version of the BioVision software originally developed by Leon Pastawa (scump1). The software performs image-based analysis of microscope images for applied research in process engineering and cultivation technology, with a focus on pellet morphology.
The current version of MicroBioVision provides tools for:
- Pellet size analysis from microscope images
- Automated image binarization and segmentation
- Extraction of morphological parameters
- Optional conversion from pixel-based measurements to micrometers for supported microscopes
The software is intended for scientific use and manual result verification.
- Python 3.8 or higher
- A working pip installation
- Tested on Windows systems (other platforms untested)
Note: The package list below may be incomplete or outdated. If an import error occurs when running the software, the missing package name is typically shown in the final line of the error message and can be installed manually via pip.
The following Python packages must be installed before running the software:
- numpy
- opencv-python
- pandas
- matplotlib
- scikit-image
- scipy
- tkinter (usually included with standard Python installations)
- PySide6
Verify Python and pip
python --version
pip --versionInstall required packages
pip install [package_name]for example
pip install numpyfor batch installation use
pip install numpy opencv-python pandas matplotlib scikit-image scipyIf multiple Python versions are installed, use:
python -m pip install numpy opencv-python pandas matplotlib scikit-image scipyNote: If installation issues occur, restarting the terminal or Python environment may resolve the problem.
- Start the software.
- Click Project in the upper left corner.
- Select Image Analysis.
- Choose the desired algorithm. (Currently, only one algorithm is available.)
- Select the images to be analysed using the upper right selection box.
- The software can reliably process up to 50 images per run.
- Exceeding this number may lead to memory-related errors.
- Restarting the software after each batch is recommended.
- Images are automatically converted to a binary representation.
- The threshold is chosen automatically but can be adjusted manually.
- Binary images are displayed alongside numerical results for manual validation.
- Pixel-to-micrometer conversion is available if the microscope magnification is provided.
- This feature currently only supports images acquired with the EVOS XL Core Imaging System.
- Results are always additionally provided in pixel units (indicated by [px]).
- For images acquired with other microscopes:
- Ignore micrometer values.
- Convert pixel-based values manually using the appropriate scale.
- To apply identical settings to all images, select: Apply Current Settings to All
- Click Analyse to start the image analysis.
- Processing time depends on the number of selected images.
- Output includes:
- Numerical result tables
- Corresponding binary images for verification
- Data can be saved using Save and Export in the bottom corner.
- Saving does not occur automatically.
- The save function depends on the active view:
- Image view: saves all binary images as .png
- Result list view: saves all result tables as .csv
- Each save operation exports all images or all tables, respectively.
- Added selection of microscope magnification
- Automatic scale conversion for EVOS XL Core Imaging System images
- Pixel-based results available for manual unit conversion
- Images from other microscopes supported without scale conversion
- Additional morphological parameters implemented:
- Maximum Feret diameter (Feret_max)
- Real pellet perimeter
- Bug fixes
- Spell checking and text corrections
- updated readme
- updated postprocessing
Note: Plans to be remembered