search

m-tassano / fastdvdnet

FastDVDnet: A Very Fast Deep Video Denoising algorithm
MIT License
586 stars 125 forks source link

A question about test #27

Closed etoilestar closed 3 years ago

etoilestar commented 3 years ago

dear author, thank you for your nice work, can you tell me how to denoise my own real images?

m-tassano commented 3 years ago

Hi, thanks for your comment.

Regarding your question, first, keep in mind that the model provided was trained for Gaussian noise. If you'd like to use it with other type of noise, it'd be better to retrain the model with the same noise distribution.

In any case, you can still use the model provided with your images, and use the noisestd input as a sort of "slider" to tune the denoising power. As for the input sequence, I would modify line #104 in _testfastdvdnet.py by removing the addition of noise seqn = seq + noise

hope this helps. best