From d55aef2a617bf8177407fe77b6899c2291e6c9e6 Mon Sep 17 00:00:00 2001 From: NBouteldja <40466985+NBouteldja@users.noreply.github.com> Date: Fri, 10 Feb 2023 14:23:48 +0100 Subject: [PATCH] Update README.md --- README.md | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index cf1da0f..793bdb2 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,9 @@ # KidneyStainTranslation -This repository provides a framework to train CycleGAN- and U-GAT-IT-based translators for unsupervised stain-to-stain translation in histology. It builds upon this [CycleGAN repository](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix). <br> -An exemplary translation from the aSMA to PAS stain in kidney tissue is provided below and its applicability by a prior segmentation model is shown by comparison with the ground-truth. The employed segmentation model is from this [paper](https://jasn.asnjournals.org/content/32/1/52.abstract) and its code repo can be found [here](https://github.com/NBouteldja/KidneySegmentation_Histology). In summary, its based on the [U-net architecture](https://arxiv.org/pdf/1505.04597.pdf) and was trained to segment several renal structures including tubulus (colorful), glomerulus , glomerular tuft , vein (including renal pelvis) , artery , and arterial lumen  from kidney histopathology data. +STAIN TRANSLATION: This repository provides a framework to train CycleGAN- and U-GAT-IT-based translators for unsupervised stain-to-stain translation in histology. It builds upon this [CycleGAN repository](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix). <br> +An exemplary translation from the aSMA to PAS stain in kidney tissue is provided below and its applicability by a prior segmentation model is shown by comparison with the ground-truth. The employed segmentation model is from this [paper](https://jasn.asnjournals.org/content/32/1/52.abstract) and its code repo can be found [here](https://github.com/NBouteldja/KidneySegmentation_Histology). In summary, its based on the [U-net architecture](https://arxiv.org/pdf/1505.04597.pdf) and was trained to segment several renal structures including tubulus (colorful), glomerulus , glomerular tuft , vein (including renal pelvis) , artery , and arterial lumen  from kidney histopathology data.<br> +Note:<br> +STAIN AUGMENTATION: You can use your trained translators to augment an annotated data set and retrain the supervised model. We applied this to our previously published CNN ([source code](https://github.com/NBouteldja/KidneyStainTranslation)) for kidney structure segmentation on PAS-stained tissue to make it stain independent. <br> +IMAGE REGISTRATION: You can also perform image registration between consecutive WSIs of different stains for stain-independent segmentation. We used the [ANHIR challenge winning registration method](https://arxiv.org/abs/2106.13150) from Fraunhofer Mevis. Please request access [there](https://www.mevis.fraunhofer.de/en/employees/johannes-lotz.html). <br> | Input aSMA image | Fake PAS translation | |:--:|:--:| -- GitLab