StickyPillars: Robust and Efficient Feature Matching on Point Clouds using Graph Neural Networks

Kai Fischer, Martin Simon, Florian Ölsner, Horst-Michael Gross, Patrick Mäder

Abstract:

Robust point cloud registration in real-time is an impor-
tant prerequisite for many mapping and localization algo-
rithms. Traditional methods like ICP tend to fail without
good initialization, insufficient overlap or in the presence of
dynamic objects. Modern deep learning based registration
approaches present much better results, but suffer from a
heavy runtime. We overcome these drawbacks by introduc-
ing StickyPillars, a fast, accurate and extremely robust deep
middle-end 3D feature matching method on point clouds.
It uses graph neural networks and performs context aggre-
gation on sparse 3D key-points with the aid of transformer
based multi-head self and cross-attention. The network out-
put is used as the cost for an optimal transport problem
whose solution yields the final matching probabilities. The
system does not rely on hand crafted feature descriptors or
heuristic matching strategies. We present state-of-art art ac-
curacy results on the registration problem demonstrated on
the KITTI dataset while being four times faster then leading
deep methods. Furthermore, we integrate our matching sys-
tem into a LiDAR odometry pipeline yielding most accurate
results on the KITTI odometry dataset. Finally, we demon-
strate robustness on KITTI odometry. Our method remains
stable in accuracy where state-of-the-art procedures fail on
frame drops and higher speeds.