Open LeeHaoRanRan opened 3 months ago
Another question is, if you don’t know in advance whether the input is normal, how to make a dataset based on the given label, instead of simply processing all 1s for exceptions and all 0s for normal ones. The dataset I use is BraTS2021
ReContrast reconstruct multi-level features of the encoder, instead of the image.
Hello, the result obtained by the decoder is a list. Has the original image been reconstructed? After I output the element "shape" in the list, it seems to be different from the "shape" of the original image.