prs-eth/Dynamic-LiDAR-Resimulation

Coding Style

Closed this issue · 1 comments

Hi,

Thanks for the great work.

I'm interested in re-implementing your project, but got confused a little bit about the coding style of the open-sourced code. It would be great if you can help me check out ideas.

In your implementation, can I say only the training part follow the nerf-studio style? I mean, all other parts including the dataset preparation, the testing, and the visualization are all not using the nerf-studio style. Is this just because following the nerfstudio style is clumsy sometimes or is there any other critical reasons that may undermine the correctness of the implementation?

Hi, Thank you for your interest in our work. Indeed only the training part "partially" follows the nerf-studio style, because our work is LiDAR based and the native nerf-studio framework is Camera based hence incompatible in lots of areas e.g. data manager, visualization. You can freely re-implement the code in other styles, I don't see any critical reason why nerf-studio has to be chosen, yet the training framework of nerf-studio is indeed very handy.