search

MaximeVandegar / Papers-in-100-Lines-of-Code

Implementation of papers in 100 lines of code.
MIT License
643 stars 94 forks source link

Regrading to "NeRF_Representing_Scenes_as_Neural_Radiance_Fields_for_View_Synthesis" #2

Closed chenmingxiang110 closed 1 year ago

chenmingxiang110 commented 1 year ago

Hi Max,

Thanks very much for your detailed tutorial on Youtube. But I have one thing pretty confused about. Taking the training dataset as an example, where the shape is 16,000,000 by 9, I see the first three numbers are the ray_origins, the mid three are the ray_directions, and the last three are the RGB colors. I am wondering how the ray origins and ray directions are computed. The problem may be tedious, but I think it would very helpful if it could be included in the video or the code.

MaximeVandegar commented 1 year ago

Hi @chenmingxiang110,

Thank you for your suggestion. The aim of this repository is to provide i) short, and ii) self-contained implementations. I did not include the ray computation in order to keep the code short, but I agree it alters the self-contained criterion.

I think the ray computation can be incorporated in 15-20 lines of code, I will try to add them in the near future, thank you.

In case you need more details about the ray computation, here is a nice script, or my Udemy course where we discuss it in details.

chenmingxiang110 commented 1 year ago

Thank you very much for your reply:-)