search

andreped / livermask

💥 Command line tool for automatic liver parenchyma and liver vessel segmentation in CT using a pretrained deep learning model
https://pypi.org/project/livermask/
MIT License
63 stars 11 forks source link

Are there performance metrics for the model? #33

Closed hossam-zaki closed 1 year ago

hossam-zaki commented 1 year ago

Describe the solution you'd like Would just like a section in the ReadMe about how well this model performs in segmenting liver parenchyma and vessels. Will be useful in integrating it into scientific papers.

andreped commented 1 year ago

Hello, @hossam-zaki!

We presented performance metrics on both liver parenchyma and vessels in the Supporting Information of this paper. See Table E in Appendix S4 in the Supplementary Information (click this link to download the PDF and see Table E).

I have included performance metrics below (which were reported on a local dataset against manual annotations):

Class DSC HD95
Parenchyma 0.946±0.046 10.122±11.032
Vessels 0.355±0.090 24.872±5.161
andreped commented 1 year ago

As this could be of interest to others, I have now updated the README to include this table: https://github.com/andreped/livermask#segmentation-performance-metrics

And I have added how to cite the performance of this tool as well: https://github.com/andreped/livermask#acknowledgements

Thank you for the great suggestion, @hossam-zaki! And best of luck with your research paper! :]

hossam-zaki commented 1 year ago

Thank you so much! Much appreciated!!