release

MxIF and H&E image registration tool

This tool is developed to establishing cell-level correspondence for cells from MxIF and H&E.

Install dependencies

pip install stardist # if detect cells on the fly
pip install tifffile
pip install probreg==0.3.7
pip install opencv-python
pip install numpy pandas matplotlib 

Apply to a case

Start from MxIF and H&E cell segmentation results

Require the tsv file exported from QuPath. The tsv file saves the coordinates and staining features of the cells from cell segmentation. Our workflow does not rely on a specific cell segmentation model. You can use any model as long as the segmentation results looks good to you.
The original MxIF and H&E images (must be tiff file) were also needed to provide extra information (for example, the size of the target image) or read in the original image and save the transformed image. You will also need specify a folder to save the transformed image. Please refer to code cli.py

python cli.py -s /opt/HE.tiff.cell_quant.tsv -t /opt/MxIF.tiff.cell_quant.tsv -is /opt/HE.tiff  -t /opt/MxIF.tiff -o /opt/temp

Directly start from MxIF and H&E image

Different from the previous one. This client code includes the cell segmentation part in the workflow, which means the client takes less input, but it takes longer time to run. Please refer to code cli_seg.py

Run with CPD (coherent point drift)

python cli_seg.py -is /opt/HE.tiff  -it /opt/MxIF.tiff -o /opt/temp

Run with CPD+GM (coherent point drift and graph matching)

python cli_seg_plus.py -s /opt/HE.tiff  -t /opt/MxIF.tiff -o /opt/temp