secdetect

Small package for segmenting serial sections from light microscopy images

Installation

1. Create conda environment (assumes anaconda or miniconda distribution already installed)

$ conda create -n secdetect numpy scikit-image shapely
$ conda activate secdetect

2. Install secdetect from github repo

$ (secdetect) pip install git+https://github.com/hoogenboom-group/secdetect.git

3. (Optional) Vastly overcomplicated but highly recommended environment setup with conda

$ (secdetect) conda install -c conda-forge matplotlib jupyterlab nodejs=15
$ (secdetect) pip install tqdm ipympl ipywidgets
$ (secdetect) jupyter labextension install @jupyter-widgets/jupyterlab-manager
$ (secdetect) jupyter labextension install jupyter-matplotlib
$ (secdetect) jupyter nbextension enable --py widgetsnbextension

Getting started

See secdetect/notebooks/example.ipynb

Other resources for image segmentation

• scikit-image segmentation examples

  • Watershed segmentation
  • Comparison of segmentation and superpixel algorithms
  • Segmenting human cells in mitosis

• ilastik

  • Download
  • Object classification workflow

• DigitalSreeni's youtube channel

  • Image segmentation using K-means
  • Image segmentation using traditional machine learning
  • Image segmentation using U-net