Adversarial point set registration

TL;DR

This paper introduces a novel one-shot adversarial learning approach for point set registration that does not require point correspondences, leveraging a critic network to align source and target point clouds effectively.

Contribution

The paper proposes a new adversarial registration method that treats point clouds as probability distributions and does not depend on point correspondences, unlike traditional algorithms.

Findings

Effective on challenging benchmarks

Outperforms existing baselines

No reliance on point correspondences

Abstract

We present a novel approach to point set registration which is based on one-shot adversarial learning. The idea of the algorithm is inspired by recent successes of generative adversarial networks. Treating the point clouds as three-dimensional probability distributions, we develop a one-shot adversarial optimization procedure, in which we train a critic neural network to distinguish between source and target point sets, while simultaneously learning the parameters of the transformation to trick the critic into confusing the points. In contrast to most existing algorithms for point set registration, ours does not rely on any correspondences between the point clouds. We demonstrate the performance of the algorithm on several challenging benchmarks and compare it to the existing baselines.