Challenging the Universal Representation of Deep Models for 3D Point Cloud Registration
This work addresses the challenge of universal representations in 3D point cloud registration, providing a baseline and dataset for further research, but it is incremental as it builds on existing methods.
The authors tackled the problem of universal representation learning for 3D point cloud registration by proposing a non-learning baseline method that either outperforms or matches state-of-the-art learning-based methods, and introduced a dataset where learning methods struggle to generalize.
Learning universal representations across different applications domain is an open research problem. In fact, finding universal architecture within the same application but across different types of datasets is still unsolved problem too, especially in applications involving processing 3D point clouds. In this work we experimentally test several state-of-the-art learning-based methods for 3D point cloud registration against the proposed non-learning baseline registration method. The proposed method either outperforms or achieves comparable results w.r.t. learning based methods. In addition, we propose a dataset on which learning based methods have a hard time to generalize. Our proposed method and dataset, along with the provided experiments, can be used in further research in studying effective solutions for universal representations. Our source code is available at: github.com/DavidBoja/greedy-grid-search.