Roni-Lab / PPSNet

PPSNet: Leveraging Near-Field Lighting for Monocular Depth Estimation from Endoscopy Videos (ECCV, 2024)

Home Page:https://ppsnet.github.io/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Can you open source your LightDepth implementation?

isJHan opened this issue · comments

It's a wonderful job. Thanks for your work!

You mentioned you implemented LightDepth on your own in your paper.

Can you release your implementation?

Thanks!
Have a nice day!

Hi @isJHan,

Thanks for your interest! Note that even with our implementation of LightDepth, there are some differences described in appendix F of our paper. Unfortunately the authors have not, and as far as I know will not, release the code publicly.

The current plan is, depending on final feedback from a conference submission, we will release our full training implementation not only for our approach, PPSNet, but approaches that we compare against such as LightDepth. We will also release pre-trained models for those other baselines at that time. It is a non-trivial effort to organize and release those codes, and additionally we want to make sure this work is accepted at a conference before open-sourcing all of our hard work for free. Please keep an eye on this repo in the meantime, I'm hoping I can do the above release sometime in mid-to-late July or so based on the current feedback timelines.

Note that our contribution, PPSNet, already has publicly available model files, loss functions, and pre-trained models available in this repo. Please let me know if you have any trouble using any of those components, I'm happy to try and help as much as I can even before some of the aforementioned, planned updates to this repo.

Best regards,

Akshay

Thanks for your reply. I will keep tracking this repo.