Marvelous work with imperfect code style

[leonardodalinky]

Recently I'm working on implementing FPS algorithms in single-thread CPU program.

This work shows an excellent performance compared to the majority of other implementations.

However, I inspect all of the codes and notice that the code style is terrible. It's generally hard to port the code to other applications directly. For example:

• Hard-coded float type for coordinates.
• Hard-coded dimension as 3.
• Abuse of macros.
• Not found any of C++17 features, even though the paper has mentioned that C++17 is used.

I have revised most of the codes with C++17 Standard to my Python Package fpsample with several benchmarks. I think my revised code in https://github.com/leonardodalinky/fpsample/tree/main/src/bucket_fps/_ext may help to refactor.

[hanm2019]

Thanks for the revision. I review the code, and it works better indeed!

[leonardodalinky]

THX. And to my surprise, is there another work of you called "FuseFPS"? Maybe someday I should port it too when the code release is available. I'll close by saying thank you for your outstanding work. :)

[hanm2019]

Yes, FuseFPS is an optimized work for QuickFPS, but It is just for the domain-special accelerator. The fusion of kd-tree construction and point sampling may cause low performance on the CPU.

[leonardodalinky]

Hi han. I wonder if you could join us to maintain the QuickFPS implementation in fpsample. I may not be familiar with the logic of your code when something goes wrong. I hope you can provide some help.

[hanm2019]

No problem, what can i do for you?

[leonardodalinky]

I have invited you the repos and the torch-fpsample v1.0 project. It would be great if you could help me finish some of the tasks in the project. Much thanks!!!