Isn't the model sensor-independent?

[study-never-alone]

Hi： Thanks for this great work. In the paper you said: Since our proposed denoising network needs to be trained for specific sensors. So, we need train sifferent model for different sensor? But, in my opinion. Eqn. (10) indicates that the model can be used for any sensor after k-Sigma Transform

[bigeagle]

Yes, we can use k-sigma transform to use the same model for multiple sensors. But the distribution of k, sigma, or the noise profile, are still "defined" by the training data, i.e. the model cannot generalize to untrained noise profiles. So we cannot claim that this method is sensor-independent.

[study-never-alone]

Thanks for your reply. What do you mean by noise profile? I think the final noise model is only related to f(x*). Can I have a contact information from you, if it's convenient, so that I can consult you in more detail?