m-tassano/fastdvdnet

Can this model denoise the video taken by the camera in real time? Or is it only possible to denoise a full video?

Closed this issue · 1 comments

Can this model denoise the video taken by the camera in real time?
Or is it only possible to denoise a full video?

As discussed in the paper, the time to denoise a frame of resolution 960 × 540 in the current implementation is about 100ms. That's only 10fps, so I'd say that'd be your main limiting factor to denoise in real time.