A denoised Mean Teacher for domain adaptive point cloud registration

TL;DR

This paper introduces a denoised Mean Teacher approach for domain adaptive point cloud registration in medical imaging, improving accuracy and robustness by filtering pseudo labels and synthesizing noise-free training pairs, especially for lung vessel registration.

Contribution

It proposes a novel denoised teacher-student framework with two denoising strategies to enhance domain adaptation in point cloud registration, outperforming existing methods.

Findings

Achieved 13.5/62.8% improvement over baseline in registration accuracy.

Set a new state-of-the-art TRE of 2.31mm on lung vessel data.

Effectively mitigated pseudo label noise in unsupervised domain adaptation.

Abstract

Point cloud-based medical registration promises increased computational efficiency, robustness to intensity shifts, and anonymity preservation but is limited by the inefficacy of unsupervised learning with similarity metrics. Supervised training on synthetic deformations is an alternative but, in turn, suffers from the domain gap to the real domain. In this work, we aim to tackle this gap through domain adaptation. Self-training with the Mean Teacher is an established approach to this problem but is impaired by the inherent noise of the pseudo labels from the teacher. As a remedy, we present a denoised teacher-student paradigm for point cloud registration, comprising two complementary denoising strategies. First, we propose to filter pseudo labels based on the Chamfer distances of teacher and student registrations, thus preventing detrimental supervision by the teacher. Second, we make the teacher dynamically synthesize novel training pairs with noise-free labels by warping its moving inputs with the predicted deformations. Evaluation is performed for inhale-to-exhale registration of lung vessel trees on the public PVT dataset under two domain shifts. Our method surpasses the baseline Mean Teacher by 13.5/62.8%, consistently outperforms diverse competitors, and sets a new state-of-the-art accuracy (TRE=2.31mm). Code is available at https://github.com/multimodallearning/denoised_mt_pcd_reg.