Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Can you provide D2-Net preprocessed images? #10

Open
doublepoints opened this issue Apr 4, 2023 · 14 comments
Open

Can you provide D2-Net preprocessed images? #10

doublepoints opened this issue Apr 4, 2023 · 14 comments

Comments

@doublepoints
Copy link

Hi, Authors,
Thank you for your time.
As the title, can you provide D2-Net preprocessed images?
As I have confirmed that this data has been removed from Loftr.
Best regards

@TruongKhang
Copy link
Owner

Hi @doublepoints,

I've just checked and found that the authors of D2-Net removed the preprocessed images. Instead, Users have to preprocess themselves here by using COLMAP.

The preprocessed images stored in my machine are huge (more than 200GB). Unfortunately, I don't have enough online space to upload and share it!

Let me know if you have any suggestions!

@Penge666
Copy link

Penge666 commented Apr 7, 2023

May I ask if the author processes the data under the Win system or Linux system? Now I'm encountering some minor difficulties in processing data.Thank you very much!

@TruongKhang
Copy link
Owner

@Penge666 , I think you should use Linux OS (ubuntu is recommended). Which data preprocessing step do you have trouble with?

@Penge666
Copy link

Penge666 commented Apr 7, 2023

Thank you for your reply. I downloaded the COLMAP execution file on the Win system and then executed the command. The ddl cannot be found. Do you want to ask for advice? bug like this(由于找不到glog.dll,无法继续执行代码。重新安装程序可能会解决此问题。)May I ask if the author could introduce the process of processing the dataset at that time.thank you very much!

@TruongKhang
Copy link
Owner

I recommend you install Ubuntu to experiment, not only for using COLMAP, but it's also beneficial when you want to use other libraries or open-source codes.
For the error above, I think you have some problems with the COLMAP installation. If the binary file you downloaded didn't work, try to install from source here

@Penge666
Copy link

Penge666 commented Apr 7, 2023

Okay, I'll try again, thank you. Excuse me,May I ask if COLMAP is processed on this dataset?

image

But the following error was reported

image

@TruongKhang
Copy link
Owner

which COLMAP command are you running?
Sorry but I don't understand your intention!

@Penge666
Copy link

Penge666 commented Apr 7, 2023

okay, thank you very much. I'll try again

@Penge666
Copy link

Excuse me, I have encountered a minor issue while processing the dataset. May I ask if COLMAP has encountered such a bug on the raw data? May I ask how the results obtained from running are correct?thank you!

image

@Penge666
Copy link

Sorry to bother you, but I would like to inquire about the content of the train folder in Magedepth.thank you very much!

@TruongKhang
Copy link
Owner

TruongKhang commented Apr 13, 2023

For the error above, I think to run python undistort_reconstruction.py, you have to download MegaDepth SfM models from https://www.cs.cornell.edu/projects/megadepth/ as well. I saw this in here

@Penge666
Copy link

Yes, I downloaded the MegaDepth SfM models, but I don't know if it's correct after processing. May I ask if the author can display the Magedepth train folder?

@TruongKhang
Copy link
Owner

in the training folder, you need two folders:

  • phoenix: this is downloaded directly from the MegaDepth website.
  • Undistorted_SfM: this folder is the one you need to produce by COLMAP. It looks like this
    image

@Penge666
Copy link

thank you very much!
This is the result of my handling

image

But these folders are empty and there are no files inside. I'm not sure if this is the case on your end

image
image
image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants