search

leekanggeun / ISCL

Official Tensorflow implementation of ISCL (Under review)
8 stars 1 forks source link

How to synthesize film and charge noise? #2

Open vfcerexwn opened 1 year ago

leekanggeun commented 1 year ago

Please check this paper "Removing Imaging Artifacts in Electron Microscopy using an Asymmetrically Cyclic Adversarial Network without Paired Training Data" with an attached figure. A4FF6F60-EF0E-49CA-8CDC-AF43B1F4687A

vfcerexwn commented 1 year ago

Thank you for your prompt reply, but I still don't understand some parts of the paper, I hope you can give me some guidance. My questions are mainly about the following three points.

  1. I read in this paper "Removing Imaging Artifacts in Electron Microscopy using an Asymmetrically Cyclic Adversarial Network without Paired Training Data" that the synthetic noise-corrupted images were not used for training the proposed model. But did you use the synthetic images to train the model?And I can't find the testing folder,only 96 images for training and 32 images for cross validation.

  2. I don't quiet understand why you choose to half the training images since that the training set contains only 96 images.

  3. I carefully read "Removing Imaging Artifacts in Electron Microscopy using an Asymmetrically Cyclic Adversarial Network without Paired Training Data", but they didn't share the code of film and charge noise model and they didn't share the training set.

At 2022-10-15 21:19:44, "leekanggeun" @.***> wrote:

Please check this paper "Removing Imaging Artifacts in Electron Microscopy using an Asymmetrically Cyclic Adversarial Network without Paired Training Data" with an attached figure.

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