Vertical Line in the output

BoomStarcuc / GeneSegNet

MIT License

14 stars 2 forks source link

Vertical Line in the output #5

Open zrsamani opened 9 months ago

zrsamani commented 9 months ago

Hi, Congratulations for this amazing work. I ran the NSCLC model on nanostring NSCLC data. The output seems to have visible lines in the borders of overlapping patches. Do you have any suggestion? I tried the NSCLC model on hippocampus data and it did not have such a problem. Do you have any suggestion?

roderickslieker commented 3 months ago

@zrsamani I am testing this as well and ran into the same observation. Did you check the original stainings? It seems to be that some parts of the image are stitched in the wrong place (in our data I am sure this is happening). I tested two fields of view for CosMx data and this occurred only in one of the two FOVs. Otherwise: great method, and really cool to use both the stainings and transcription data, although the documentation is still a bit sparse...