root-tissue-seg-package 1.0.6

An mlf-core prediction package for root tissue segmentation.

Root-Tissue-Segmentation Package

Prediction package for reproducible U-Net models, trained for semantic segmentation of microscopy images of root tissue from A. thaliana (https://github.com/qbic-pipelines/root-tissue-segmentation-core/). These models are trained using the mlf-core framework and tested for reproducibility. This package can be deployed within an analysis pipeline as a module for root tissue segmentation (rts) of fluorescence microscopy images.

• Free software: MIT

• Documentation: https://rts-package.readthedocs.io.

Package Tools

• Segmentation prediction CLI: rts-pred

• Uncertainty of prediction CLI: rts-pred-uncert

• Input feature importance (Guided Grad-CAM) CLI: rts-feat-imp

Usage Examples

• rts-pred -i ./brightfields -o ./predictions -m mark1-PHDFM-u2net-model.ckpt --suffix ""

• rts-pred-uncert -i ./brightfields -o ./predictions -m mark1-PHDFM-u2net-model.ckpt --suffix "" -t 5

• rts-feat-imp -i ./brightfields -o ./predictions -m mark1-PHDFM-u2net-model.ckpt --suffix "" -t 2

Credits

This package was created with mlf-core using cookiecutter.

Changelog

This project adheres to Semantic Versioning.

1.0.0 (2021-08-10)

Added

Fixed

Dependencies

Deprecated

0.1.0 (2021-08-10)

Added

• Created the project using mlf-core

Fixed

Dependencies

Deprecated