2021/03/14 by Xin Wen, Wen, Xin, Zhizhong Han +10 · 2 citations
Computer Science · Engineering · Mathematics · #3D Shape Modeling and Analysis #Advanced Numerical Analysis Techniques #Algorithm #Artificial intelligence #Coding (social sciences) #Computer Graphics and Visualization Techniques #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Consistency (knowledge bases) #Domain (mathematical analysis) #FOS: Computer and information sciences #Focus (optics) #Mathematics #Point cloud #Rigid transformation #Transformation (genetics) #cs.CV
paper · pdf · doi:10.48550/arxiv.2103.07838
Accepted to CVPR 2021
openalex publication_date 2021/03/14 · openalex created_date 2021/03/29 · arxiv created 2021/06/12 · arxiv updated 2021/06/15 · openalex updated_date 2026/08/08
In this paper, we present a novel unpaired point cloud completion network, named Cycle4Completion, to infer the complete geometries from a partial 3D object. Previous unpaired completion methods merely focus on the learning of geometric correspondence from incomplete shapes to complete shapes, and ignore the learning in the reverse direction, which makes them suffer from low completion accuracy due to the limited 3D shape understanding ability. To address this problem, we propose two simultaneous cycle transformations between the latent spaces of complete shapes and incomplete ones. The insight of cycle transformation is to promote networks to understand 3D shapes by learning to generate complete or incomplete shapes from their complementary ones. Specifically, the first cycle transforms shapes from incomplete domain to complete domain, and then projects them back to the incomplete domain. This process learns the geometric characteristic of complete shapes, and maintains the shape consistency between the complete prediction and the incomplete input. Similarly, the inverse cycle transformation starts from complete domain to incomplete domain, and goes back to complete domain to learn the characteristic of incomplete shapes. We provide a comprehensive evaluation in experiments, which shows that our model with the learned bidirectional geometry correspondence outperforms state-of-the-art unpaired completion methods.