TIO-IKIM / CellViT

CellViT: Vision Transformers for Precise Cell Segmentation and Classification
https://doi.org/10.1016/j.media.2024.103143
Other
217 stars 33 forks source link

Training and inference transformations not found #13

Closed huangzhii closed 1 year ago

huangzhii commented 1 year ago

Hello!

Thanks for deploying this great repo. I was trying to find training and inference transformations. However, I haven't been able to locate the details, such as mean and standard deviation. Consequently, this has made it challenging for me to execute the code, including obtaining the necessary self.run_conf["transformations"] data.

Could you please provide the detail of your training and inference transformations?

FabianHoerst commented 1 year ago

Hello,

the transformations are described in the appendix in the paper. There you should find every entry.

huangzhii commented 1 year ago

Thank you. It works perfectly. I evaluated manually and this is indeed the best nuclei segmentation algorithm so far.