PuzzleNet: Boundary-Aware Feature Matching for Non-Overlapping 3D Point Clouds Assembly

We address the 3D shape assembly of multiple geometric pieces without overlaps, a scenario often encountered in 3D shape design, field archeology, and robotics. Existing methods depend on strong assumptions on the number of shape pieces and coherent geometry or semantics of shape pieces. Despite rai...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of computer science and technology Ročník 38; číslo 3; s. 492 - 509
Hlavní autoři: Liu, Hao-Yu, Guo, Jian-Wei, Jiang, Hai-Yong, Liu, Yan-Chao, Zhang, Xiao-Peng, Yan, Dong-Ming
Médium: Journal Article
Jazyk:angličtina
Vydáno: Singapore Springer Nature Singapore 01.06.2023
Springer
Springer Nature B.V
Témata:
Algorithms
Archaeology
Artificial Intelligence
Assembly
Automation
Computer Science
Data Structures and Information Theory
Datasets
Information Systems Applications (incl.Internet)
Iterative methods
Manufacturing engineering
Neural networks
Regular Paper
Robotics
Semantics
Software Engineering
Theory of Computation
Three dimensional models
point cloud
shape assembly
3D registration
boundary feature
geometric learning
ISSN:1000-9000, 1860-4749
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:We address the 3D shape assembly of multiple geometric pieces without overlaps, a scenario often encountered in 3D shape design, field archeology, and robotics. Existing methods depend on strong assumptions on the number of shape pieces and coherent geometry or semantics of shape pieces. Despite raising attention to 3D registration with complex or low overlapping patterns, few methods consider shape assembly with rare overlaps. To address this problem, we present a novel framework inspired by solving puzzles, named PuzzleNet, which conducts multi-task learning by leveraging both 3D alignment and boundary information. Specifically, we design an end-to-end neural network based on a point cloud transformer with two-way branches for estimating rigid transformation and predicting boundaries simultaneously. The framework is then naturally extended to reassemble multiple pieces into a full shape by using an iterative greedy approach based on the distance between each pair of candidate-matched pieces. To train and evaluate PuzzleNet, we construct two datasets, named DublinPuzzle and ModelPuzzle, based on a real-world urban scan dataset (DublinCity) and a synthetic CAD dataset (ModelNet40) respectively. Experiments demonstrate our effectiveness in solving 3D shape assembly for multiple pieces with arbitrary geometry and inconsistent semantics. Our method surpasses state-of-the-art algorithms by more than 10 times in rotation metrics and four times in translation metrics.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1000-9000
1860-4749
DOI:10.1007/s11390-023-3127-8