Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Learning Iterative Robust Transformation Synchronization

About

Transformation Synchronization is the problem of recovering absolute transformations from a given set of pairwise relative motions. Despite its usefulness, the problem remains challenging due to the influences from noisy and outlier relative motions, and the difficulty to model analytically and suppress them with high fidelity. In this work, we avoid handcrafting robust loss functions, and propose to use graph neural networks (GNNs) to learn transformation synchronization. Unlike previous works which use complicated multi-stage pipelines, we use an iterative approach where each step consists of a single weight-shared message passing layer that refines the absolute poses from the previous iteration by predicting an incremental update in the tangent space. To reduce the influence of outliers, the messages are weighted before aggregation. Our iterative approach alleviates the need for an explicit initialization step and performs well with identity initial poses. Although our approach is simple, we show that it performs favorably against existing handcrafted and learned synchronization methods through experiments on both SO(3) and SE(3) synchronization.

Zi Jian Yew, Gim Hee Lee• 2021

Related benchmarks

TaskDatasetResultRank
Point cloud registration3DMatch (test)
Registration Recall84.6
339
Rigid Registration3DLoMatch (test)
RR76.8
43
Point cloud registrationETH--
38
3D Point Cloud Registration3DMatch
Translation Error (cm)12
20
Multiview RegistrationScanNet 30 scans 18
RE@3°54.3
19
Multiway point cloud registration3DLoMatch
Rotation Error (°)11.52
16
Multiway point cloud registrationKITTI
RE (°)5.17
16
Multi-view RegistrationScanNet (test)
Rotation Error (< 3°)47.4
15
Multiway point cloud registrationNSS
RR (%)65.9
8
Multiway point cloud registrationNSS
RE (Deg)11.42
8
Showing 10 of 11 rows

Other info

Follow for update