icvteam / pgnet Goto Github PK

tkank you for you nice work. could you tell us what evaluation code do you use? please

Dear author, may I ask if it is convenient to disclose the code about the Boundary Displacement Error?

Thank you for your nice work.I noticed in the paper that S3 and S4 of the Swin-Transformer branch have the same size and are not downsampled. What is the reason?

Thank you for your nice work. I had some problems with the code, my model predictions were all grey with a little bit of texture. What might be the cause of the error?

Could UHRSD dataset provide links to Baidu Web disk

A total of 5,920 images were manually selected from websites (e.g. Flickr Pixabay) with free copyright.

Could you list the licenses of each individual image and where they are from exactly? I checked a few images, but could not find where it says "free copyright", whatever that means.

• 000056.jpg https://www.flickr.com/photos/193480849@N05/51323576150/ "All rights reserved"
• 000111.jpg https://www.flickr.com/photos/mikehousephotography/51289703416/ "All rights reserved"
• 001539.jpg https://www.flickr.com/photos/pjvmartinsphotography/51867151692/ "All rights reserved" (Not exactly the same photo, but I think it is also there somewhere)

Thank you for your nice work. I would like to know how to visualize the feature maps，like figure 4. Thank you very much.

Hi Team this is really great job can you please let me know where to download the trained models.

HI team,

Thanks for the checkpoint I have saved the folder PGNet_DUT%2BHR but getting the following error

thank you for your nice work. I want to run the train_distributed.py file. but it is not work. the dataset.py seem to not ready. could you tell me the problems.

Hey

Great work! well done!
The google drive link contains the model that was trained on DUTS + HR, named Ours-DH.
Can you please share the Ours-UH model as well?