Dataset loading problem

[ZhouChen-start]

Hello, I have downloaded the data from https://zenodo.org/records/10066853, but it seems to be the original image and cannot be used directly. I am very sorry, I am very interested in this project, can you provide me with some data processing methods? Thank you very much. If there are friends who see this problem, I also hope you can help me, thank you.

[pushpak-pati]

Thank you for the query.
The preprocessing code is now included in the repo at https://github.com/AI4SCR/VirtualMultiplexer/tree/master/preprocess

[firstme2024]

During my reading, I have some questions about the following issues and would appreciate your guidance:

Question 1: When HE generates IHC (CD44, NKX3.1, AR, CD146, p53), the real IHC is randomly masked, and the training verifies whether it can generate complete and reliable IHC. since we need to divide the WSI image into multiple patches, the IHC is randomly masked; which method is it:

(1) First WSI is randomly masked first, and then divided into multiple patches；

(2) First WSI is divided into multiple patches, and then each patch is randomly masked.

[ZhouChen-start]

Thank you for the query. The preprocessing code is now included in the repo at https://github.com/AI4SCR/VirtualMultiplexer/tree/master/preprocess

Thank you very much for your reply. Now I have another question. The image downloaded from https://zenodo.org/records/10066853 is named H4_TMA_15_02_IVB_HE.png, but the name in HE_splits.csv is H4_TMA_15_02_IVB_HE_95.24064_2.2. I don’t understand the meaning of the final suffix 95.24064_2.2. Can you tell me?