{ "id": 555808, "title": "Denoising and Raw WBP Tomos", "url": "/competitions/czii-cryo-et-object-identification/discussion/555808", "author_name": "SimWdm", "post_date": "2025-01-09T09:52:25.699000", "votes": 1, "comment_count": 0, "views": 0, "content": "

Dear Competition Organizers,

\n

Thank you for organizing this challenge and for providing carefully annotated experimental tomograms — those are fantastic contributions to the community!

\n

I have two questions regarding the training and test data:

\n
    \n
  1. Which denoiser was applied to the raw WBP tomograms?
    2. \n
  2. Would it be possible to provide the raw WBP tomograms for the test cases? (If I understand correctly, right now we only have access to denoised tomograms at test time).
    3. \n
\n

I am asking because I would like to experiment with pre-training. Most available datasets are not denoised, so I would need to either apply a denoiser similar to the one used for the competition data or work directly with raw WBP tomograms for both training and testing.

\n

I noticed that other participants have also expressed interest in accessing raw tomograms and learning more about the denoising process (see here).

\n

Best,
\nSimon

", "messages": [ { "id": 3092217, "postDate": "2025-01-09T09:52:25.700Z", "content": "

Dear Competition Organizers,

\n

Thank you for organizing this challenge and for providing carefully annotated experimental tomograms — those are fantastic contributions to the community!

\n

I have two questions regarding the training and test data:

\n
    \n
  1. Which denoiser was applied to the raw WBP tomograms?
    2. \n
  2. Would it be possible to provide the raw WBP tomograms for the test cases? (If I understand correctly, right now we only have access to denoised tomograms at test time).
    3. \n
\n

I am asking because I would like to experiment with pre-training. Most available datasets are not denoised, so I would need to either apply a denoiser similar to the one used for the competition data or work directly with raw WBP tomograms for both training and testing.

\n

I noticed that other participants have also expressed interest in accessing raw tomograms and learning more about the denoising process (see here).

\n

Best,
\nSimon

", "rawMarkdown": "Dear Competition Organizers,\n\nThank you for organizing this challenge and for providing carefully annotated experimental tomograms — those are fantastic contributions to the community!\n\nI have two questions regarding the training and test data:\n\n1. Which denoiser was applied to the raw WBP tomograms?\n2. Would it be possible to provide the raw WBP tomograms for the test cases? (If I understand correctly, right now we only have access to denoised tomograms at test time). \n\nI am asking because I would like to experiment with pre-training. Most available datasets are not denoised, so I would need to either apply a denoiser similar to the one used for the competition data or work directly with raw WBP tomograms for both training and testing.\n\nI noticed that other participants have also expressed interest in accessing raw tomograms and learning more about the denoising process (see [here](https://www.kaggle.com/competitions/czii-cryo-et-object-identification/discussion/555247#3089572)).\n\nBest,\nSimon", "votes": 1 } ], "comments": [], "raw_markdown_by_id": { "3092217": "Dear Competition Organizers,\n\nThank you for organizing this challenge and for providing carefully annotated experimental tomograms — those are fantastic contributions to the community!\n\nI have two questions regarding the training and test data:\n\n1. Which denoiser was applied to the raw WBP tomograms?\n2. Would it be possible to provide the raw WBP tomograms for the test cases? (If I understand correctly, right now we only have access to denoised tomograms at test time). \n\nI am asking because I would like to experiment with pre-training. Most available datasets are not denoised, so I would need to either apply a denoiser similar to the one used for the competition data or work directly with raw WBP tomograms for both training and testing.\n\nI noticed that other participants have also expressed interest in accessing raw tomograms and learning more about the denoising process (see [here](https://www.kaggle.com/competitions/czii-cryo-et-object-identification/discussion/555247#3089572)).\n\nBest,\nSimon" } }