Morphing and Sampling Network for Dense Point Cloud Completion

By Minghua Liu, Lu Sheng, Sheng Yang, Jing Shao, Shimin Hu. AAAI Conference on Artificial Intelligence, 2019

375 citations recorded by the source corpus.

3D point cloud completion, the task of inferring the complete geometric shape from a partial point cloud, has been attracting attention in the community. For acquiring high-fidelity dense point clouds and avoiding uneven distribution, blurred details, or structural loss of existing methods' results, we propose a novel approach to complete the partial point cloud in two stages. Specifically, in the first stage, the approach predicts a complete but coarse-grained point cloud with a collection of parametric surface elements. Then, in the second stage, it merges the coarse-grained prediction with the input point cloud by a novel sampling algorithm. Our method utilizes a joint loss function to guide the distribution of the points. Extensive experiments verify the effectiveness of our method and demonstrate that it outperforms the existing methods in both the Earth Mover's Distance (EMD) and the Chamfer Distance (CD).

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Environmental Engineering

View Paper: https://www.semanticscholar.org/paper/f0ff1e0739494fcbaeefc3e30038457ed606b1c4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Yilun Du
2. Jiajun Wu
3. Jiwen Lu
4. Jie Zhou
5. Yongming Rao
6. Zexiang Xu
7. Linqi Zhou
8. Zhizhong Han

In Computer Graphics and Computer-Aided Design

1. Yilun Du
2. Zexiang Xu
3. Hao Su
4. Minghua Liu
5. Bo Dai
6. Jiwen Lu
7. Chen Change Loy
8. Linqi Zhou

In Environmental Engineering

1. Hao Su
2. Jiwen Lu
3. Zhizhong Han
4. Yongming Rao
5. Yu-Shen Liu
6. Jie Zhou
7. Jiageng Mao

Citations in the corpus

• One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization (2023, methodology): “To endow intelligent agents with this ability, many existing methods [30, 18, 24, 10, 81, 85, 15, 77] exploit class-specific priors by training 3D generative networks on 3D shape datasets [4].”
• GRNet: Gridding Residual Network for Dense Point Cloud Completion (2020, methodology): “To relief the structure loss caused by MLP, AtlasNet [29] and MSN [30] recover the complete point cloud of an object by estimating a collection of parametric surface elements.”
• 3D Shape Generation and Completion through Point-Voxel Diffusion (2021, background): “First, we note EMD is a better metric for measuring completion quality because by solving the linear assignment problem it forces model outputs to have the same density as the ground-truths [23] and it is known that CD is blind to visual inferiority [1].”
• Unsupervised 3D Shape Completion through GAN Inversion (2021, methodology): “We compare our PatchVariance against expansion penalty [21] and repulsion loss [35].”
• SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution with Skip-Transformer (2021, result): “Compared with the previous methods, which often ignore the spatial relationship among points [56, 43, 31] or simply learn through self-attention in a single level of multi-step point cloud decoding [53, 29, 46], our skip-transformer is proposed to integrate the spatial relationships across different”
• PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers (2021, methodology): “CD-(cid:96) 2 ( × 1000) AtlasNet [13] PCN [51] FoldingNet [50] TopNet [37] MSN [17] NSFA [52] PFNet [16] CRN [43] GRNet [48] between the encoder and decoder (model B).”