Hello, sorry to bother you, I have looked up a lot of data, I have a big question about the noise estimation network, is it generating a fixed value mask? I would like to know how he has a guiding role for the network? How does the noise level here affect denoising? I really don't understand this question. After visualizing, I found that after the last tanh function, the values inside were all 1, so I was very confused and did not know if I made a mistake somewhere. Looking forward to your reply! Thanks a lot!
Hello, sorry to bother you, I have looked up a lot of data, I have a big question about the noise estimation network, is it generating a fixed value mask? I would like to know how he has a guiding role for the network? How does the noise level here affect denoising? I really don't understand this question. After visualizing, I found that after the last tanh function, the values inside were all 1, so I was very confused and did not know if I made a mistake somewhere. Looking forward to your reply! Thanks a lot!