lllyasviel / lllyasviel/ControlNet

Can ControlNet be used to do virtual staining of hisopathological (H&E) images?

Open
#162 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
34.1k
Forks
3k
PR merge metrics
No merged PRs in 30d

Description

I was wondering if ControlNet could be used to virtually stain an unstained image for my research purposes. For those not aware of how unstained and stained H&E images look like, they look like [this[(https://www.google.com/search?sxsrf=AJOqlzWGwEsKcZEGMowYq7gSsUmbJNS3VQ:1675370125229&q=unstained+to+stained+h%26e+image&tbm=isch&sa=X&ved=2ahUKEwj1ibmJ2Pf8AhUTUzUKHRiHCwcQ0pQJegQIDxAB&biw=1873&bih=969&dpr=1#imgrc=x3HbfcorYe-2yM). The unstained and stained images would have the same structure & texture (essentially the same segmentation mask) but different color obviously. I was wondering if with enough training images, the model could accurately stain the unstained images to make them look like real H&E stained images.

Would appreciate any input, thanks a lot in advance!

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are named. Start by reviewing ControlNet's conditioning and training documentation, then assess whether paired unstained and H&E images could support the proposed virtual-staining experiment; done means a documented feasibility answer.

Written by the indexing model from the issue text.

Assessment

Domain
computer-vision, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
15/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.