REGTR: End-to-end Point Cloud Correspondences With Transformers

Abstract

Despite recent success in incorporating learning into point cloud registration, many works focus on learning feature descriptors and continue to rely on nearest-neighbor feature matching and outlier filtering through RANSAC to obtain the final set of correspondences for pose estimation. In this work, we conjecture that attention mechanisms can replace the role of explicit feature matching and RANSAC, and thus propose an end-to-end framework to directly predict the final set of correspondences. We use a network architecture consisting primarily of transformer layers containing self and cross attentions, and train it to predict the probability each point lies in the overlapping region and its corresponding position in the other point cloud. The required rigid transformation can then be estimated directly from the predicted correspondences without further post-processing. Despite its simplicity, our approach achieves state-of-the-art performance on 3DMatch and ModelNet benchmarks. Our sou

Stats

• citations236
• S2 citations—
• github stars267
• HF likes0
• heat score22.67
• arxiv keyyew2022regtr

Related papers

• Transmatcher: Deep Image Matching Through Transformers For Generalizable Person Re-identification (2021)4.68
• Coe: Deep Coupled Embedding For Non-rigid Point Cloud Correspondences (2024)0.00
• Training Vision Transformers For Image Retrieval (2021)0.00
• Lahnet: Local Attentive Hashing Network For Point Cloud Registration (2025)0.00
• Improving The Matching Of Deformable Objects By Learning To Detect Keypoints (2023)7.74
• Attention-based Multimodal Image Matching (2021)8.60
• End-to-end Learning Of Keypoint Detector And Descriptor For Pose Invariant 3D Matching (2018)12.40
• Latformer: Locality-aware Point-view Fusion Transformer For 3D Shape Recognition (2021)6.34