Copy link Copy Markdown

Contributor

No description provided.

Copy link Copy Markdown

Contributor

Thanks. Do we still want to support CovarianceType templating to retain the performance of reprojection error?

Copy link Copy Markdown

Contributor Author

Thanks. Do we still want to support CovarianceType templating to retain the performance of reprojection error?

I think we discussed that the overhead is minimal? We can easily bring it back, if needed. The initial concern was that this would affect the performance of the unweighted cost functions but this is not the case anymore now.

Closed

ahojnnes deleted the user/jsch/cov-weighted-cost-functor branch

November 3, 2024 16:40

Comment on lines +126 to +127

		ReprojErrorConstantPoseCostFunctor(const Eigen::Vector2d& point2D,
		const Rigid3d& cam_from_world)

Copy link Copy Markdown

Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ahojnnes What justifies this reordering?

Copy link Copy Markdown

Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Without this change, the point2D_cov and point2D parameters are separated by the cam_from_world in the weighted covariance wrapper (both in C++ and Python).

Merged