Dear Community,

I want to perform automated segmentation of histological whole slide images (WSIs) of kidney tissue to extract morphological features associated with various structural units like glomerulus, proximal tubules, arteries etc. I came across a couple of articles which demonstrate automated segmentation of kidney tissue using convolutional neural network (CNN) based packages like U-net, nn-Unet, etc. However, they do not provide the pre-trained model (with the weights) or the labeled images that were used for training the models (ground truth data).

Hence, I am reaching out to see if there is a work around which allows me to perform automated segmentation of kidney tissue without having to train a CNN model from scratch and not use labeled images annotated by trained pathologists. In case, there are studies/articles which have already addressed this issue please feel free to share them. Hope this will also help researchers who find themselves in a similar situation and are looking to find a solution. I really appreciate your help!

Regards