weijun88 / sanet Goto Github PK

From my understanding, i have some questions:

1. Did you just use a color changed original image and not use it together with the original color image during training?
   If this is true, is it enough for the model to focus more on visual content?
2. Did you just transfer the colors of the images in the train set without using anything from the test set?
   If this is true, could you please explain your strategy?
   Thank you.

Hi, can you give the WEIGHT of this model please, thanks!

Hi,
you have done a splendid work for Polyp Segmentation, and I have a question about PCS. In section 3.3, you proposed a Probability Correction Strategy to explicitly correct the predicted probability through logit reweighting.
However, the logit of positive samples(i.e., logit > 0) is still greater than 0 after the normalization, and the result after sigmoid is greater than 0.5, so it is still recognized as a positive sample. I cannot understand your intention to use the PCS.

It would be very kind of you to answer my question.
Thanks.

I follow the settings of the readme, the code has not been modified in any way, but the loss has not decreased. Is this normal?

can u share SFA's code?