Coarse-to-Fine Point Cloud Registration with SE(3)-Equivariant Representations

Lin, Cheng‐Wei; Chen, Tung-I; Lee, Hsin-Ying; Wen-Chin, Chen,; Hsu, Winston H.

doi:10.48550/arxiv.2210.02045

Cited by 2 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Notably, some works extend the original representation by integrating SO(3) or SE(3) equivariance for tasks such as reconstruction [8], point cloud registration [48,19], and manipulation [35]. Zhu et al [48] learned SO(3)-equivariant features to perform correspondence-free point cloud registration, while Lin et al [19] used SE(3)-equivariant representations to obtain and refine the registration result globally and locally. Simeonov et al [35] learned SE(3)-equivariant object representations for manipulation and estimated relative transforms through optimization.…”

Section: B Neural Implicit Representations For Roboticsmentioning

confidence: 99%

“…Simeonov et al [35] learned SE(3)-equivariant object representations for manipulation and estimated relative transforms through optimization. These methods target point clouds known to be associated with the same object and can suffer from performance degradation for partially overlapped point clouds [48,19] or require iterative refinement to recover the desired relative transform [35].…”

Section: B Neural Implicit Representations For Roboticsmentioning

confidence: 99%

See 1 more Smart Citation

NeuSE: Neural SE(3)-Equivariant Embedding for Consistent Spatial Understanding with Objects

Fu,

Du,

Singh

et al. 2023

Robotics: Science and Systems XIX

View full text Add to dashboard Cite

We present NeuSE, a novel Neural SE(3)-Equivariant Embedding for objects, and illustrate how it supports object SLAM for consistent spatial understanding with long-term scene changes. NeuSE is a set of latent object embeddings created from partial object observations. It serves as a compact point cloud surrogate for complete object models, encoding full shape information while transforming SE(3)-equivariantly in tandem with the object in the physical world. With NeuSE, relative frame transforms can be directly derived from inferred latent codes. Our proposed SLAM paradigm, using NeuSE for object shape and pose characterization, can operate independently or in conjunction with typical SLAM systems. It directly infers SE(3) camera pose constraints that are compatible with general SLAM pose graph optimization, while also maintaining a lightweight object-centric map that adapts to real-world changes. Our approach is evaluated on synthetic and real-world sequences featuring changed objects and shows improved localization accuracy and change-aware mapping capability, when working either standalone or jointly with a common SLAM pipeline.

show abstract

Section: B Neural Implicit Representations For Roboticsmentioning

confidence: 99%

Section: B Neural Implicit Representations For Roboticsmentioning

confidence: 99%

NeuSE: Neural SE(3)-Equivariant Embedding for Consistent Spatial Understanding with Objects

Fu,

Du,

Singh

et al. 2023

Robotics: Science and Systems XIX

View full text Add to dashboard Cite

show abstract

“…Notably, some works extend the original representation by integrating SO(3) or SE(3) equivariance for tasks such as reconstruction [8], point cloud registration [47,19], and manipulation [35]. Zhu et al [47] learned SO(3)-equivariant features to perform correspondence-free point cloud registration, while Lin et al [19] used SE(3)-equivariant representations to obtain and refine the registration result globally and locally. Simeonov et al [35] learned SE(3)-equivariant object representations for manipulation and estimated relative transforms through optimization.…”

Section: B Neural Implicit Representations For Roboticsmentioning

confidence: 99%

“…Simeonov et al [35] learned SE(3)-equivariant object representations for manipulation and estimated relative transforms through optimization. These methods target point clouds known to be associated with the same object and can suffer from performance degradation for partially overlapped point clouds [47,19] or require iterative refinement to recover the desired relative transform [35].…”

Section: B Neural Implicit Representations For Roboticsmentioning

confidence: 99%