YeastInference

Installation

These instructions have only been tested on Ubuntu and Macintosh, but should also work on other operating systems.

1. Install Miniconda with python 3.7

2. Create a new conda environment and activate it:

   conda create -n yeast-env python
   conda activate yeast-env 
   

3. Install pytorch-cpu:

   On Windows & Linux:

   conda install pytorch torchvision cpuonly -c pytorch
   

   On Mac:

   conda install pytorch torchvision -c pytorch
   

4. Clone this repo and install dependencies:

   git clone https://github.com/imagirom/YeastInference.git
   cd YeastInference
   pip install -r requirements.txt
   pip install -e git+https://github.com/inferno-pytorch/inferno@cache-reject#egg=inferno-0.3.1

Usage

Place the TIF files you want to process in a new folder YeastInference/data. In the YeastInference folder, run

python predict_rois.py

The predictions should be generated as zip files next to the input images.

To learn about further options, please run python predict_rois.py -h.