Question: Rationale for using rays_o instead of rays_o × rays_d for Plucker embedding's first three dimensions

[sustecher]

Hi @Robbyant,

I noticed in the implementation of Plücker embedding at cam_utils.py#L115, the first three dimensions use rays_o directly.
In contrast, a more common/general practice for Plücker embedding (as seen in CamI2V/base/base.py#L156) uses the cross product of rays_o and rays_d (rays_o × rays_d) for these dimensions.

I'm curious to understand the design decision here:

1. What is the rationale for using rays_o instead of rays_o × rays_d in this implementation?
2. Are there specific use cases/scenarios where this choice is more appropriate for the lingbot-world project?

Thanks for your time and clarification!

[taylover-pei]

I have the same doubt. Could you please help clarify it? Thank you.

[taylover-pei]

@EzioBy