English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present Point-TTA, a novel test-time adaptation framework for point cloudregistration (PCR) that improves the generalization and the performance ofregistration models. While learning-based approaches have achieved impressiveprogress, generalization to unknown testing environments remains a majorchallenge due to the variations in 3D scans. Existing methods typically train ageneric model and the same trained model is applied on each instance duringtesting. This could be sub-optimal since it is difficult for the same model tohandle all the variations during testing. In this paper, we propose a test-timeadaptation approach for PCR. Our model can adapt to unseen distributions attest-time without requiring any prior knowledge of the test data. Concretely,we design three self-supervised auxiliary tasks that are optimized jointly withthe primary PCR task. Given a test instance, we adapt our model using theseauxiliary tasks and the updated model is used to perform the inference. Duringtraining, our model is trained using a meta-auxiliary learning approach, suchthat the adapted model via auxiliary tasks improves the accuracy of the primarytask. Experimental results demonstrate the effectiveness of our approach inimproving generalization of point cloud registration and outperforming otherstate-of-the-art approaches.

Point-TTA: Test-Time Adaptation for Point Cloud Registration Using Multitask Meta-Auxiliary Learning