sai-bi / L1Flattening

An L1 Image Transform for Edge-Preserving Smoothing and Scene-Level Intrinsic Decomposition
MIT License
67 stars 17 forks source link

An L1 Image Transform for Edge Preserving Smoothing and Scene Level Intrinsic Decomposition #1

Open RoyiAvital opened 8 years ago

RoyiAvital commented 8 years ago

Hi, You mention it is different than the article. Could you say what is different and why?

Thank You.

sai-bi commented 8 years ago

Generally it is some modifications related to image transforms as well as parameters. But the algorithms stay the same.

RoyiAvital commented 8 years ago

I see. Are results on the Paper produced using the current code or the one in the Paper?

By the way, your University page with the PDF isn't available anymore.

sai-bi commented 8 years ago

The result in the original paper is not produced with this code. The original code is in C++ and it is too messy. But they generally produce the same result.

I have left the university now. Currently I haven't built a new personal webpage.

tnestmeyer commented 8 years ago

So is the "Scene Level Intrinsic Decomposition" (Sec. 4 of your paper) part of this code? I would like to reproduce the intrinsic images results, but from the demo file I only understand how to get the "Image Flattening" step in Fig. 7, not the full pipeline. If it is implemented, could you give me a hint, what I have to define as param?

tnestmeyer commented 7 years ago

I would like to get back to my last question. Could you please elaborate on that? I plan to evaluate your intrinsic decomposition method on another dataset. Therefore, I need the full intrinsic decomposition pipeline of your method. Could you help me with that? Thanks!

ttsesm commented 6 years ago

@tnestmeyer I noticed in your paper that you are comparing with this approach. Did you manage to find the code for the whole pipeline including the reflectance/shading estimation? And if yes can I get it from somewhere.

tnestmeyer commented 6 years ago

@TheodoreT Sorry, can't help you there. In the end I used the results available for IIW at the public project page, so I did not evaluate the method on any other datasets.

RoyiAvital commented 6 years ago

@tnestmeyer , Something is wrong with the README.md file. Things aren't shown properly.

tnestmeyer commented 6 years ago

@RoyiAvital You mean in this repo, correct? I saw that myself, too, but I am neither the Owner of, nor related to this project. You probably want to address @soundsilence for fixing it.

ttsesm commented 6 years ago

@tnestmeyer I see. Thanks for the feedback.