SmartPath is an open-source multimodal whole slide histopathology imaging system with the capabilities of automaticmodality switching, region of interest detection, and run-time image processing, empowered by deep learning. The acquisition program is written in Jupyter Notebook/JupyterLab and uses Pycro-Manager for hardware control.
Install anaconda/miniconda.
Open terminal or Anaconda Prompt in Windows.
Install required packages:
$ conda env create --name smartpath --file env.yml
$ conda activate smartpath
$ pip install pycromanager
Install PyTorch, pick GPU or CPU version.
Micro-manager 2.0 gamma, OpenScan, QuPath and FIJI. See document for details.
See step-by-step instructions in the notebook.
Supervised and weakly supervised classifier for histological datasets: https://github.com/uw-loci/histo-classifier.
Run-time image enhancement for laser scanning microscopy with self-supervised denoising and single-image super-resolution: https://github.com/uw-loci/lsm-run-time-enhancement.