Example image and Model for virtual histological staining of unlabeled autopsy tissue

Example images can be found at "Example_images.zip". Here's what you will find:

Input: Autofluorescence images of unstained autopsy tissue sections; each FOV corresponds to a .mat file, with the 1st channel representing the DAPI channel and the 2nd channel representing the TxRed channel.
Output: Virtually stained H&E images generated by the network model.
Ground Truth: Histochemically stained H&E images for reference/comparison.
10 individual FOVs are provided as example images. Each FOV measures 2048 x 2048 pixels, corresponding to roughly 333 x 333 µm². The 10 FOVs are labeled from "001" to "010".

"001"-"005": These correspond to tissue regions suffering from severe autolysis, rendering staining artifacts in the histochemically stained images (ground truth).
"006"-"010": These correspond to well-preserved tissue regions with high histochemical staining quality.
How to use the model file and the corresponding codes can be found at: https://github.com/liyuzhu1998/Autopsy-Virtual-Staining/tree/main