search

borglab / gtsam

GTSAM is a library of C++ classes that implement smoothing and mapping (SAM) in robotics and vision, using factor graphs and Bayes networks as the underlying computing paradigm rather than sparse matrices.
http://gtsam.org
Other
2.53k stars 753 forks source link

Avoid dense fill-in during margianlization #1586

Open sunbingfeng opened 1 year ago

sunbingfeng commented 1 year ago

Feature

Update the marginalizeLeaves function to avoid dense fill-in.

Motivation

marginalizeLeaves will add new marginalization factor, which may cause the pose graph to be more and more dense.

The original iSAM package provided a sparsification method based on Chow-liu tree, but the tree construction is very time consuming, especially when the >20 keys are involved.

So, I'm wondering if GTSAM will provide some alternative method to sparsify the pose graph.

Thanks!

Pitch

marginalizeLeaves will keep the pose graph sparse.

dellaert commented 1 year ago

No plans. Happy to consider a PR :-)