CODAvision

CODAvision is an open-source Python package designed for microanatomical tissue labeling. It facilitates deep learning-based semantic segmentation of Whole Slide Images (WSI) through a user-friendly interface.

Quick Install

1. Download and Install Miniconda

   Follow the instructions here to download and install Miniconda.

2. Create and Activate the CODAvision Environment

   Open your terminal or command prompt and run the following commands:

   conda create -n CODAvision python=3.9.19
   
   conda activate CODAvision

3. Install CUDA Toolkit and cuDNN

   CODAvision requires the CUDA Toolkit and cuDNN to run on NVIDIA GPUs. Follow the instructions here to install the CUDA drivers. After installing the drivers, install the CUDA Toolkit and cuDNN using conda:

   conda install -c conda-forge cudatoolkit=11.2 cudnn=8.1.0

4. Install CODAvision

   Install the CODAvision package using pip:

   pip install -e git+https://github.com/Kiemen-Lab/CODAvision.git#egg=CODAvision

Warning: You might need to install Git from the following link to be able to run the pip install git link command.

Warning: After installing the package, restart your IDE and activate the environment again using conda activate CODAvision to ensure all the dependencies are properly loaded.