Cell Descriptor Lib

The Cell Descriptor Library serves as a tool designed to streamline the analysis of diverse descriptors, including mask descriptors and texture descriptors, within 2D+t datasets of cells. Examples of such datasets can be found at Cell Tracking Challenge (http://celltrackingchallenge.net/). This project was developed during PV162 course at FI MUNI.

Components

Before installing any packages, we recommend using a virtual environment:

python -m venv env
./env/Scripts/activate

To install required packages and the lib it self, run:

pip install -r .\requirements.txt
pip install .

DescriptorLib

• descriptors.py - module contains library of various descriptors
• descriptor_provider - contains list of all implemented decriptors

Exporter

Command line tool for running the computation on the dataset. Example usage

python .\src\DescriptorLibUtils\export.py .\tests\testdata\images\ .\tests\testdata\masks\

Data explorer

Module for simple parsing of the directory structure. First, inicialize the module:

from DescriptorLibUtils import DataExplorer
de = DataExplorer("../output/")

And then you can simply use objects method to explore the data:

desc = de.GetDescriptorsForCell(frame_number,cell_id)["Mask descriptors"]
imgs, masks = de.GetCellTimeline(cell_id)
mask_timeline = de.GetCellDescriptorTimeline(cell_id, "Mask descriptors")