3DPCTN: Two 3D Local-Object Point-Cloud-Completion Transformer Networks Based on Self-Attention and Multi-Resolution

Acquiring high-fidelity 3D models from real-world scans is challenging. Existing shape completion-methods are incapable of generating details of objects or learning complex point dis-tributions. To address this problem, we propose two transformer-based point-cloud-completion networks and a coarse-to...

Full description

Bibliographic Details
Main Authors: Cheng, Y. (Author), Huang, S. (Author), Li, H. (Author), Shi, Y. (Author), Tan, J. (Author), Yang, Z. (Author)
Format: Article
Language:English
Published: MDPI 2022
Subjects:
Online Access:View Fulltext in Publisher