부산대학교 도서관

We address the 3D shape assembly of multiple geometric pieces without overlaps,a scenario often encoun-tered 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 com-plex 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 leverag-ing 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 ap-proach based on the distance between each pair of candidate-matched pieces.To train and evaluate PuzzleNet,we con-struct 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 assem-bly for multiple pieces with arbitrary geometry and inconsistent semantics.Our method surpasses state-of-the-art algo-rithms by more than 10 times in rotation metrics and four times in translation metrics.