search

akshaydudhane16 / Burstormer

SOTA for Burst Super-resolution, Low-light Burst Image Enhancement, Burst Image De-noising
119 stars 5 forks source link

Question for Cyclic Burst Sampling in NRFE. #6

Open cvkaiming opened 1 year ago

cvkaiming commented 1 year ago

I saw the Cyclic Burst Sampling module in your article, and I think this module is very cleverly designed, but I can't find the definition of the code for this module in your code, can I ask you where this code is? There is one more question, the weights in the trace 1 part you gave me don't seem to match the model, is the weight file uploaded wrongly? image thank you

shekshaa commented 1 year ago

Here is the answer to your first question: there is a class called no_ref_back_projection in Burst De-noising/Network.py that performs the cyclic burst sampling in the forward pass. It uses torch.roll to shift the burst and then reshape both the original and shifted versions to couple the frames and finally concatenate the results.

cvkaiming commented 1 year ago

thanks for your replying. ---- Replied Message ---- | From | @.> | | Date | 07/21/2023 04:02 | | To | akshaydudhane16/Burstormer @.> | | Cc | seeuao @.>, Author @.> | | Subject | Re: [akshaydudhane16/Burstormer] Question for Cyclic Burst Sampling in NRFE. (Issue #6) |

Here is the answer to your first question: there is a class called no_ref_back_projection in Burst De-noising/Network.py that performs the cyclic burst sampling in the forward pass. It uses torch.roll to shift the burst and then reshape both the original and shifted versions to couple the frames and finally concatenate the results.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>