automorph's Introduction

AutoMorph 2022 👀

--Code for AutoMorph: Automated Retinal Vascular Morphology Quantification via a Deep Learning Pipeline.

Please contact [email protected] or [email protected] if you have questions.

Project website: https://rmaphoh.github.io/projects/automorph.html

Talks on NIHR Moorfields BRC: https://moorfieldsbrc.nihr.ac.uk/case-study/research-report/

News 👀

2023-08-24 update: Added feature measurement for disc-centred images; removed unused files in M3 folders.

Pixel resolution

The units for vessel average width, disc/cup height and width, and calibre metrics are defined as microns. For it, we need to organise a resolution_information.csv which includes the pixel resolution information, which can be queried in FDA or Dicom files. Alternatively, some people use approximate value for every images, e.g., 0.008 for Topcon 3D-OCT.

If you don't use these features or care their units, you can just run following command after putting all images in the folder of ./images

python generate_resolution.py

Running AutoMorph

Running with Colab

Use the Google Colab and a free Tesla T4 gpu Colab link click.

Running on local/virtual machine

Install and use on your own machines LOCAL.md.

Running with Docker

Zero experience in Docker? No worries DOCKER.md.

Common questions

Memory/ram error

We use Tesla T4 (16Gb) and 32vCPUs (120Gb). When you meet memory/ram issue in running, try to decrease batch size:

./M1_Retinal_Image_quality_EyePACS/test_outside.sh -b=64 to smaller, e.g., 32 or 16.
./M2_Artery_vein/test_outside.sh --batch-size=8 to smaller
./M2_lwnet_disc_cup/test_outside.sh --batchsize=8 to smaller

Invalid results

In csv files, invalid values (e.g., optic disc segmentation failure) are indicated with -1.

Components

Vessel segmentation BF-Net
Image pre-processing EyeQ
Optic disc segmentation lwnet
Feature measurement retipy

Citation

@article{zhou2022automorph,
  title={AutoMorph: Automated Retinal Vascular Morphology Quantification Via a Deep Learning Pipeline},
  author={Zhou, Yukun and Wagner, Siegfried K and Chia, Mark A and Zhao, An and Xu, Moucheng and Struyven, Robbert and Alexander, Daniel C and Keane, Pearse A and others},
  journal={Translational vision science \& technology},
  volume={11},
  number={7},
  pages={12--12},
  year={2022},
  publisher={The Association for Research in Vision and Ophthalmology}
}

automorph's People

Contributors

Stargazers

Watchers

automorph's Issues

Units of calculated values

Thank you for all your contributions to this library! Could you please provide a list of units for each value calculated by the tool? I believe this would be highly beneficial for other researchers referencing your work.

Some questions about resolution_information

I am very grateful that your program has been a great help to our work. The fundus imaging device we use has a resolution of 6 nm, but after I change the resolution information in resolution_information.csv, it seems to have no effect on the feature measurement results.

Thank you very much for sharing these codes. Very helpful. I am running it on Google Colab and it works OK on 45 degree fundus photographs.

I know you would not recommend using it on ultra-wide field images (Optos), but I would love to see how it performance on them. Would the input be the RGB image or channels need to be split and green fed in?

Also, in terms of the output csvs, I noticed that Disc_Zone_B_Measurement.csv has exact same values as Disc_Zone_C_Measurement.csv in 'Results\M3\Disc_centred' directory. Can I fix that somehow?

Originally posted by @lajos-cs in #1 (comment)

not enough values to unpack in tortuosity_measures.evaluate_window

Hi,
first of all, I really appreciate your work.
In the M3_feature_zone/retipy/create_datasets.py, you expect the function - tortuosity_measures.evaluate_window will return 15 values, but it only produces 13 values. Is it supposed to be like that? What am I missing?

Thanks,
Shvat

Recommend Projects

rmaphoh / automorph Goto Github PK