How can model and data files be accessed?

[LashaO]

Hey! I appreciate your effort. I am working on an open-source tree detection model - it could be worth it to include your work. The paper pointed to this repository for data and model weights, but I can't seem to find them. What would be the best place to look? Thanks

[cwinkelmann]

Dear @sulenurtopgull,
thank you for your work and the paper. I would be very interested in the dataset as well.

Thanks a lot.

[sulenurtopgull]

Dear @cwinkelmann ,
Thank you for your interest! The dataset is now available in the repository. Please check the README file for the link. Let me know if you need any further information. 🙂

[sulenurtopgull]

Thanks @LashaO for your interest in our work. I have now added the dataset to the repository, and you can find the link in the README file. As for the model weights, we are in the process of uploading them and will make all of them available soon. However, the model weights for the first 8 experiments are already included in the repository. Let me know if you have any questions! 🚀

[LashaO]

Thanks much @sulenurtopgull !

[cwinkelmann]

@sulenurtopgull Thank you very much for sharing the data. I have had a look at it. If I get the numbers right, you selected 218 Full HD images but removed all but kept only 45 because which were not suitable for the dataset? On what basis did you do that? Isn't introducing this a bias into the dataset since you seemingly selected only the simplest to detect trees? Tiles from the 45 tiles were filtered to again to remove difficult image tiles, right?

I am asking because in the survey I am analysing I have big issues with false positives on empty images or images with difficult image quality.

Aside from that, in the readme seems to be a typo for Exp-4, mAP 0.50 can not be 0.534

Best

[Maria2025S]

Hello,
thank you very much for your work,
I would like to ask if your code is available for open accees.
best

[roozbeh442]

Hi @sulenurtopgull,
Thank you for your great work. I'm having issues download the dataset and the model weights. above comments suggest that a link has been added to the Readme file, but I'm unable to find it. are you able to advise?
thanks
Rooz

[mohabtarek96]

@LashaO I can't find the dataset link, were you able to find and download it? if so, can you please send it to me?

[bw4sz]

Just letting those in this thread know that we are developing a large benchmark wrapping the many datasets out there. https://milliontrees.idtrees.org/

[bw4sz]

The dataset link is in the README.md history but was removed

• Download the dataset here !

Can you confirm why and what the intention was?

[mohab-soliman-96]

@bw4sz IDK I found the link in the history too, but it asks me for access.