Abstract
Three-dimensional point cloud registration (PCReg) has a wide range of applications in computer vision, 3D reconstruction and medical fields. Although numerous advances have been achieved in the field of point cloud registration in recent years, large-scale rigid transformation is a problem that most algorithms still cannot effectively handle. To solve this problem, we propose a point cloud registration method based on learning and transform-invariant features (TIF-Reg). Our algorithm includes four modules, which are the transform-invariant feature extraction module, deep feature embedding module, corresponding point generation module and decoupled singular value decomposition (SVD) module. In the transform-invariant feature extraction module, we design TIF in SE(3) (which means the 3D rigid transformation space) which contains a triangular feature and local density feature for points. It fully exploits the transformation invariance of point clouds, making the algorithm highly robust to rigid transformation. The deep feature embedding module embeds TIF into a high-dimension space using a deep neural network, further improving the expression ability of features. The corresponding point cloud is generated using an attention mechanism in the corresponding point generation module, and the final transformation for registration is calculated in the decoupled SVD module. In an experiment, we first train and evaluate the TIF-Reg method on the ModelNet40 dataset. The results show that our method keeps the root mean squared error (RMSE) of rotation within 0.5 and the RMSE of translation error close to 0 m, even when the rotation is up to [−180, 180] or the translation is up to [−20 m, 20 m]. We also test the generalization of our method on the TUM3D dataset using the model trained on Modelnet40. The results show that our method’s errors are close to the experimental results on Modelnet40, which verifies the good generalization ability of our method. All experiments prove that the proposed method is superior to state-of-the-art PCReg algorithms in terms of accuracy and complexity.
Highlights
IntroductionThe learning-based Point cloud registration (PCReg) method processes unordered point clouds and extracts features through a deep learning network [9,10,11]; the similarity of these features is used to calculate the transformation
Point cloud registration (PCReg) refers to the problem of finding the rigid transformation that maximizes the overlap between similar sections of two or more point clouds.As a fundamental technique in 3D data processing, it is employed in many fields including computer vision, robotics, medical image analysis and computer-assisted surgery.Researchers in the past have proposed methods [1,2,3,4,5] to address the PCReg problem.many of them are prone to converging to local optima
We propose the leveraging of transform-invariant features in the PCReg problem and evaluate the expressivity of the features; We propose a novel PCReg method that is robust to the large rigid transformation between source clouds and target clouds; We evaluate the performance of our method under several settings, demonstrating the effectiveness of the proposed method
Summary
The learning-based PCReg method processes unordered point clouds and extracts features through a deep learning network [9,10,11]; the similarity of these features is used to calculate the transformation. Most of these methods cannot cope with large transformations [7,12]; they achieve high accuracy only when the rotation and translation are limited to [−45◦ , 45◦ ] and [−0.5 m, 0.5 m], respectively. Since they are not robust to transformation and cannot act as stable inputs for DNNs, DNNs cannot learn the features with transformation invariance [13,14]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
