How to use U-NET Image Segmentation on Kaggle Nuclei detection Data Set?

I am trying to implement U-NET Image Segmentation on Kaggle Nuclei Detection Dataset in MATLAB.

In MATLAB, U-NET can be implemented on any imagery dataset IF we have images and their corresponding masks.

For example in below picture, we have a an image and its corresponding mask. In such a data set we can impelement U-NET segmentation.

Question:

How to implement U-NET Segmentation if we have seperated masks( a single image has multiple masks ccorresponding to each object), just like in Kaggle Nuclei Dataset.

If it has to be implement by combining all masks for an image into a single mask, then

How to calculate IoU, TP,TN,FP and FN between predicted and given image?