---
license: cc-by-nc-sa-4.0
---

# CONCH embeddings for CAMELYON16 dataset

[CONCH](https://github.com/Mahmoodlab/CONCH) is a vision-language foundation model created by the Mahmood Lab. We can use CONCH to embed patches from whole slide images.

In this dataset, we have embedded the CAMELYON16 whole slide images (n=399 slides). See this link for the CAMELYON16 dataset: https://camelyon17.grand-challenge.org/Data/ .

The vision-only directory contains embeddings with `proj_contrast=False`, and the `vision-language` directory contains embeddings with `proj_contrast=True`.

## Install

Create a conda environment with the required libraries.

```shell
conda create -n conch \
    python pytorch torchvision pytorch-cuda=11.8 transformers numpy \
    openslide openslide-python scikit-learn timm regex ftfy h5py pandas \
    -c pytorch -c conda-forge
conda activate conch
python -m pip install git+https://github.com/mahmoodlab/CONCH
```

## Get patch coordinates

Use [CLAM](https://github.com/mahmoodlab/CLAM) to extract patch coordinates from whole slide images. Jakub has modified CLAM to extract patches of a constant physical size, but he still has to upload that code to GitHub.

## Embed patches

Log in to HuggingFace Hub with the `huggingface_hub` [command line interface](https://huggingface.co/docs/huggingface_hub/en/guides/cli#huggingface-cli-login) or the [Python packge](https://huggingface.co/docs/huggingface_hub/en/package_reference/login#huggingface_hub.login). Make sure you [request access to CONCH](https://huggingface.co/MahmoodLab/CONCH) with your HuggingFace Hub account.

Embed the patches using the script `extract_features.py`. These embeddings are suitable for vision-only models, like an ABMIL model to infer a slide-level label.

```shell
python extract_features.py \
    --wsi-dir data/slides \
    --patch-dir data/patches \
    --wsi-extension '.tif' \
    --save-dir data/embeddings \
```

If you want to get the similarity of patches to text prompts, then also include the `--proj-contrast` argument.

```shell
python extract_features.py \
    --wsi-dir data/slides \
    --patch-dir data/patches \
    --wsi-extension '.tif' \
    --proj-contrast \
    --save-dir data/embeddings-projcontrast
```