mask

[zhouhao028]

Hello, thank you very much for your work! I have a question now, why do both train and val in dataset.py need mask input. Inputting the image and its mask in the training set helps the model recognize the area it needs to recognize, but why does the validation set also need the mask input? Also, if I only input the image without the mask area, can VM-Unet finally output the mask image obtained based on this image input? Thank you very much for your work, thank you very much!

[JCruan519]

Q1: The mask is needed in the validation set to calculate performance metrics.
Q2: Yes, during testing, you can input only the image and still obtain the mask.